Subscribe now

Cybersecurity teams spend considerable resources building defenses, collecting logs, and developing detection rules. Yet a question often remains unanswered: Can those systems actually recognize the attacks they are supposed to catch?

The distinction matters as enterprises introduce more cloud services, AI applications, and AI-assisted development tools. Security environments are becoming more complicated, but the growing volume of telemetry does not guarantee that critical information is being collected or interpreted correctly. Logs can be missing, pipelines can fail silently, and detection rules can overlook malicious activity.

Brava Security wants to give organizations a way to test those assumptions. The company has emerged from stealth with $15 million in funding led by F2 Venture Capital and Pitango VC. Its multi-agent AI technology generates the exact telemetry an attack would produce without actually executing it, allowing enterprises to determine which attacks their existing defenses would miss. Axios was the first to report the news.

The platform also helps address the gaps it identifies, without requiring organizations to replace their current security infrastructure.

Replacing Assumptions With Evidence

The problem Brava addresses sits beneath the growing interest in AI-powered security operations. AI agents can investigate alerts and help security teams make decisions, but they depend on the evidence generated by an organization’s existing security systems.

If an attack never triggers an alert because the relevant telemetry was not collected, an AI agent has little to investigate. Brava’s argument is that organizations need to establish the reliability of their detection infrastructure before they can confidently automate more of their security operations.

“Everyone is racing to put AI in charge of the SOC. You can’t automate a decision based on evidence no one ever proved,” said Idan Basre, co-founder and CEO of Brava Security. “We’re solving the problem that comes before AI can make the right decision: making sure the security infrastructure actually sees the threats it’s supposed to detect.”

Basre encountered the issue firsthand while leading product management for Microsoft Defender. His experience informs the company’s focus on identifying weaknesses in the security data and detection processes that enterprises already operate.

Understanding the Attack Before It Happens

Brava’s platform models attacker behavior and determines how different attack techniques would manifest across an organization’s environment. It then evaluates whether existing defenses would capture the resulting activity and identifies missing telemetry, ineffective detection rules, and other gaps.

The Security Operations platform can rewrite SIEM detections and data feeds that are blind to modeled attacks. It also investigates alongside security teams, bringing together evidence that helps them understand the weaknesses in their infrastructure.

The technology’s ability to generate telemetry without executing an attack is central to its approach. Rather than waiting for an actual incident to expose a detection failure, organizations can examine how their systems would respond to modeled attacker behavior.

“The hard part of security was never writing detections. It’s knowing exactly what an attack will look like in a specific environment, across all of its log sources, before it ever happens. That’s the problem our models were built to solve,” said Itay Gabbay, co-founder and CTO of Brava Security.

Early Adoption Across Enterprise Markets

Brava is entering the market with customers that include Fortune 100 companies across retail, energy, and hospitality. Its customer list also includes Cushman & Wakefield and WELL Health, one of Canada’s largest healthcare companies.

The company has prioritized compatibility with existing enterprise security environments. Its platform integrates with Splunk, Google SecOps, CrowdStrike, and more than 70 other sources and destinations.

Deployment speed is another part of its offering. Brava says its first insights become available within 10–15 minutes of onboarding, allowing security teams to begin evaluating their findings without a lengthy implementation process.

Cushman & Wakefield CISO Erik Hart described the platform as part of the organization’s broader effort to apply AI to security operations.

“Our SecOps journey is about applying AI where it makes our people more effective, and Brava sits at the center of that. It investigates alongside our team, bringing evidence together so decisions rest on facts rather than on searching” said Hart.

Funding the Next Stage of Growth

Brava was founded by Basre and Gabbay, whose backgrounds combine security research and AI development. Basre is a Talpiot alumnus who spent ten years in Unit 8200 as a threat researcher before leading product management for Microsoft Defender. Gabbay spent seven years in the IDF building core AI systems, including work on a joint DARPA-Israeli research program.

The company has approximately 20 employees, including AI researchers and security researchers. Its R&D operations are based in Israel, while its commercial team is expanding across the United States.

The $15 million seed round also included angel investors, among them executives from Anthropic, Microsoft, and Check Point. Brava plans to use the capital to accelerate product development and expand commercially across North America, with financial services and government among its next target markets.

For Brava, the opportunity lies in making security performance something enterprises can demonstrate rather than simply assume. As AI takes on more responsibility in the SOC, that distinction could become increasingly important.

Subscribe now

Subscribe now

The traditional boundaries of workplace cybersecurity are becoming less relevant as employees move between cloud applications, browser-based services, and AI platforms throughout the day. A single browser session can involve access to customer records, company documents, financial systems, and external applications. Protecting these interactions requires more than monitoring individual devices.

Guardz is expanding its approach to workspace security through an integration with Palo Alto Networks’ Prisma Browser for Business, as first reported by SiliconANGLE. The addition brings browser protection into Guardz’s platform for managed service providers (MSPs), connecting it with existing security capabilities across identity, email, endpoints, data, and user risk.

Connecting The Security Ecosystem

The integration addresses a growing problem for security teams: attacks frequently cross multiple environments, but the tools monitoring those environments do not always share context.

A phishing email, for example, can lead an employee to a fraudulent login page designed to steal credentials or active session tokens. Prisma Browser for Business can block the malicious destination, while Guardz correlates the browser event with the original email, related Microsoft 365 activity, and endpoint telemetry.

This allows MSPs and Guardz’s 24/7 managed detection and response (MDR) team to investigate related activity as a unified incident.

“Identity is the new center of gravity in cybersecurity, and the browser has become the primary gateway to the modern workspace. Traditional security tools operate in silos, leaving critical gaps in visibility and response,” said Dor Eisner, CEO and Co-Founder of Guardz.

Prisma Browser for Business also brings inline machine-learning phishing detection, Advanced URL Filtering, WildFire file analysis, upload and download scanning, and risky browser extension detection into the Guardz environment.

A More Accessible Deployment Model

For MSPs, integrating another security capability across multiple customer environments can introduce additional administrative work. Guardz says it has simplified provisioning, setup, and default configuration, allowing providers to deploy Prisma Browser for Business within minutes.

Kevin Kehoe of Datawise Networks, an early-access MSP, highlighted the implementation experience: “The rollout through our RMM was super smooth. Within minutes, the Prisma Browser for Business extension was on every Chrome and Edge profile, and it was quick and easy to configure and deploy.”

The integration is intended to help MSPs extend browser protection through workflows they already use, rather than requiring a separate and complicated rollout.

Shivam Srivastava, VP Product at Palo Alto Networks, said the collaboration brings browser security into the services small businesses already receive from their providers. “For most small businesses, the browser is now the office. Working with Guardz brings Prisma Browser for Business into the security these organizations receive from the providers they already trust. It connects our browser protection with the wider coverage and support their MSP delivers, helping safeguard the applications and information employees use every day,” he said.

Bringing AI Usage Into Security Operations

The integration also complements Guardz’s AI Usage and Control capabilities. As employees increasingly use browser-based AI services, organizations need ways to understand adoption and control how business information is shared.

MSP technicians can establish client-specific restrictions on copying, pasting, uploading, and downloading files by website or category. Guardz has also introduced an AI usage report that provides a per-client view of AI tool adoption.

Eisner described the broader direction behind the partnership: “Together with Palo Alto Networks, we’re advancing a new era of identity-centric security operations, connecting signals across identity, browser, email, endpoint, and data into a unified security fabric that is powered by agents. Guardz enables MSPs to autonomously investigate, correlate, and respond to threats across the entire workspace, while giving small businesses the freedom to embrace AI securely and thrive. This is the future of cybersecurity: identity-centric, unified data fabric, and built for autonomous defense at scale,” he added.

The integration reflects Guardz’s effort to move beyond isolated security controls toward a connected model that gives MSPs greater visibility across the digital environments their customers depend on.

Subscribe now

Subscribe now

Twice the pull requests. Twice the review time for each one. That is the pattern among engineers who use AI heavily, and it is the problem Hud‘s newly launched agentic rollouts are built to solve.

The numbers come from Faros AI’s 2025 vendor survey. Teams with high AI adoption merge 98% more pull requests. Their pull request review time rises 91%. Output went up. So did the time spent reviewing it.

Hud’s new capability targets that gap directly. Agentic rollouts let coding agents take a change from pull request to production, with Hud validating every step against real production behavior. The goal is to bring the speed of shipping closer to the speed of coding.

Reading the Data

The Faros AI figures describe a velocity gain that runs into a slower review stage. Engineers using AI heavily produce about twice as many pull requests. Each of those pull requests takes about twice as long to review.

A second data point covers what happens after review. Harness’s 2025 vendor survey found that only 6% of organizations describe their continuous delivery as fully automated.

Put together, the picture is clear. Agents write code fast. Review slows down. Fully automated delivery is rare. The fastest stage of engineering now feeds into the slowest, and the more code agents write, the bigger the backlog becomes. Every team that adopts AI hits this wall.

Where the Time Goes

Hud pinpoints the friction. A pull request waits for someone willing to merge it, because it is hard to trust an agent to ship without breaking production. The agent cannot see how its code runs there. Reviewers approve changes they did not have time to read. After the deploy, someone watches dashboards. A change written in minutes can wait for days. Human-driven validation stretches time to delivery.

None of those steps involves writing code. All of them involve confidence. And confidence has been built by hand, one review and one dashboard at a time.

How Hud Builds That Confidence

Hud starts with measurement. Its runtime code sensor, added with a one-line SDK, automatically learns baseline behavior at the function level in production. One line to install. It learns on its own.

From there, every change is checked at three points.

Pre-merge. The pull request is analyzed against a production impact map built on real parameters and code flows. Hud classifies its production risk level. Risky changes come with clear context that speeds review. Safe ones can merge as fast as they were written.

In flight. As the new version takes traffic, Hud compares behavior to the baseline continuously. Regressions are caught immediately. Function-level context identifies the root cause and the pull request responsible.

Post-regression. Hud’s context supports rollback, revert, and a fix to the underlying issue. The coding agent receives deep context so it does not repeat the mistake.

The outcome: teams know the production-level impact of each change before, during, and after rollout.

What Changes for the Team

The shift is from watching to knowing. Without runtime evidence, confidence comes from people: a reviewer reading the diff, an engineer watching the dashboard. With Hud, it comes from data collected at the function level. Agents and engineers look at the same evidence. Safe pull requests merge at the speed they were written. Risky ones arrive with context. Regressions are caught, traced, and reversed. The coding agent gets the details it needs to do better next time.

One Number From the Field

Riverside provides a concrete benchmark. The platform for recording and editing studio-quality audio and video ships fast with AI. Much of its backend is real-time media code that only shows its true behavior under production load.

Riverside runs every deploy through a canary with Hud’s agentic rollout and automatic rollback. Flawed deploys are detected quickly. They are reverted in about eight minutes. Usually, that happens before a customer notices.

Shilo Magen, VP of R&D at Riverside, put it this way: “Our engineers use AI agents every day, and writing code is no longer the bottleneck. The real challenge is shipping that code to production with confidence.”

“For a real-time media platform like Riverside, even a small regression can have a major impact on the user experience. With Hud, we can automatically detect regressions and roll back problematic releases in minutes. Our engineers now spend less time worrying about deployments and more time building.”

The Founder’s Read

Roee Adler, Co-Founder and CEO of Hud, connected the numbers to the product. His argument: coding and triage have agents, while the merge and rollout still run on people.

“AI already changed how code gets written. The bigger problem now is getting that code into production safely, at the speed agents write it. Every engineering leader I talk to is trying to solve this problem. Coding and triage have agents, but the merge and rollout still run on people, because it’s scary to give coding agents accountability over the stability and reliability of the production environment. That’s the missing link in the AI software factory, and it’s the problem we built Hud to solve. With agentic rollouts, agents and engineers work from the same runtime context, at the function level, so a change is merged to production based on evidence. When something does regress, it’s immediately detected and fixed, but more importantly the agent that wrote it gets a feedback loop to learn and improve.”

Hud’s direction sits in the loop itself. Every rollout teaches the coding agent how its changes behave in production. Each new change ships with less risk than the last. That feedback loop is where Hud is headed.

Subscribe now

Subscribe now

Pluto Security‘s attribution rests on evidence the operator left in the code and on a mistake that tied the accounts together. Across a cryptominer and backdoor campaign spanning 28 repositories, the language of the code comments, the choice of infrastructure, the target set and the timing of activity all point one way: a financially motivated, Chinese-speaking operator or small group.

The comments in the source

Wave-one code carries comments in Simplified Chinese that read as operational notes. Pluto Security cites “daemon disguise name” for process masquerading, “blend into Xinference directory” for the targeted environment, and “kill rival processes” for clearing competing miners. These are English renderings of comments written in Simplified Chinese.

Infrastructure that lines up

The hosting reinforces the language evidence. The wave-one command-and-control server, 8.137.187.63, sits on Alibaba Cloud. The miner binary is mirrored on ModelScope and hf-mirror.com. The targeting assumes Xinference and Tencent Cloud, and the SSH injection paths include the xinference user’s home. Pluto Security says language and hosting place the operator in the Chinese-speaking ML ecosystem.

The clock

Activity clusters between 18:38 and 21:19 UTC, which is 02:38 to 05:19 China Standard Time, late night in that zone.

The operator’s own slip

The link between accounts came from a mistake. A test account held backup copies of the loader, modeling_evil.py.bak-src and modeling_evil.py.bak-sync, in the qwen2-tiny-lite repository. They were older versions of the same malware with different download URLs. Those URLs did not point at the test account, dsxk4mr8tq. They pointed at hhzz667 on Hugging Face and wssb123e on ModelScope. The operator had copied production code into a new test account, updated the URLs in the main file and forgot the backups, so its own version history connected the accounts.

Profit as the motive

Pluto Security says the money points the same way. The payoff is Pearl and Monero wallets with real pool payouts: the first-wave wallet has received about 2,276 PRL, roughly $5,000, across 716 transactions, and its live workers report peakminer. The tooling is commodity, with throwaway accounts, a leaked token and development files named evil. Once its own miner is running, the payload also removes rival miners so that the compute goes to its own. In Pluto Security’s words, the wallets and tradecraft show the goal is profit.

One line is different. The soumikmahato “parasit” build carries Indonesian-language comments and a different build system. Pluto Security says that fits either a second operator in the same niche or one operator running a deliberately different-looking line for redundancy.

What the “evil” filename shows

In wave one, the file name modeling_evil.py turned up on the development account, dsxk4mr8tq. The wave-one production build, qwen2-final, used clean names, Qwen2Config and modeling_qwen2.py, and the EvilConfig naming only existed in the dev account. The wave-two flagship is different: MENXIANGYU/qwen2-tiny points auto_map at modeling_evil2.EvilModel and had been downloaded 18,012 times by October 25.

The development trail

The code shows a clear progression. Phase one, on September 21, was a 37-line SSH backdoor that planted a single ed25519 key in authorized_keys. Phase two, on September 21 and 22, grew to 102 lines, with key injection in six paths and .bashrc injection. Phase three, on September 22, was the 305-line fileless miner, qwen2-final. The wave-one command-and-control ran on four ports, and Pluto Security counts more than 40 code iterations across the campaign.

Pluto Security’s assessment

Pluto Security’s assessment is a financially motivated, Chinese-speaking operator or small group, based on the code comments, the hosting, the targets and the timing, with more than 21,000 downloads across the campaign and its linked models. Language and hosting place the operator in the Chinese-speaking ML ecosystem, and the wallets and tradecraft show the goal is profit.

Subscribe now

Subscribe now

A $400 million funding round at a $3.8 billion valuation puts Upwind Security among the cloud security companies attracting substantial investor attention. The company, which reported that its revenue quadrupled in 2026, is raising capital as enterprises contend with expanding cloud infrastructure, more complex application environments, and the growing use of AI.

Bessemer Venture Partners led the Series C, with Canapi Ventures and existing investors participating. The new valuation marks a significant increase from the $1.5 billion valuation attached to Upwind’s Series B in January 2026. The company said the funding will support product development and go-to-market expansion.

The announcement offers a snapshot of both Upwind’s growth and the problem it is trying to address: security teams need to make decisions across cloud environments that are constantly changing, while AI adds new forms of activity to monitor and control.

A Market Defined by Operational Complexity

Cloud security platforms must account for more than the number of assets an organization operates. Security teams also need to understand how those assets interact, which services are exposed, and whether a weakness can be reached through an actual path in the environment.

That work becomes harder when applications span multiple cloud providers and development teams make frequent changes. Security tools that produce separate findings for workloads, APIs, and other infrastructure components can leave teams to piece together the wider picture themselves.

Upwind’s platform combines runtime intelligence with context from cloud workloads, APIs, and AI agents. It is designed to help security, development, and DevOps teams understand risks in relation to live environments, prioritize remediation, and detect attacks in real time.

“AI is fundamentally changing both the cloud infrastructure architecture and the scale of cyber threats and how organizations build and operate software,” said Amiram Shachar, CEO and Co-Founder of Upwind. “Context is key to managing internal and external risks, security teams need to understand how cloud, and AI interact in real time. We built Upwind to unify those signals and eliminate complexity, helping organizations identify their most critical exposures, fix risks proactively, and detect attacks as they happen.”

From Cloud Visibility to Security Operations

The company’s platform is intended to help customers move beyond collecting findings toward deciding which issues deserve attention first. By connecting information about runtime activity with security risks, the system can help teams evaluate whether a potential exposure has practical significance and coordinate remediation accordingly.

Upwind supports Amazon Web Services, Microsoft Azure, Google Cloud Platform, and Oracle Cloud Infrastructure. Its AWS Security Hub integration gives customers access through the AWS console, with simplified onboarding and unified billing.

The company’s customer roster includes Siemens, Roku, Carvana, Nubank, Supabase, and ClickUp. Those names provide evidence of commercial reach across different sectors, although Upwind has not disclosed customer-level revenue, deployment sizes, or retention figures in the funding announcement.

The platform also incorporates AI agents that can investigate threats autonomously, validate attack-surface findings by acting as red teams, and remediate infrastructure risks at machine speed. These capabilities extend the company’s focus from visibility and prioritization into security workflows that can take action.

A Broad Investor Syndicate

Bessemer Venture Partners led the round, joined by Canapi Ventures and existing investors Cyberstarts, Greylock Partners, Craft Ventures, Leaders Fund, Salesforce Ventures, TCV, Alta Park Capital, Penny Jar Capital, Cerca Partners, Swish Ventures, and Picture Capital.

The participation of existing backers alongside the new lead investor gives Upwind additional capital to expand its product and commercial operations. The announcement does not disclose how the proceeds will be divided between research and development, hiring, sales, or international expansion.

Shachar described the financing as a way to build on the company’s market traction. “Our investors have backed our vision and the traction we’re seeing in the market,” Shachar said. “This round gives us the opportunity to accelerate our momentum, expand our capabilities, and continue delivering the innovation customers need to secure cloud and AI infrastructure in the agentic era.”

What the Next Stage Will Require

Funding and revenue growth offer useful signals about a company’s position, but they do not settle the competitive question. Cloud security vendors must show that their platforms can improve how organizations identify and resolve risk, integrate with existing infrastructure, and keep pace with new applications and AI-driven workloads.

Upwind’s reported fourfold revenue growth in 2026 points to increasing commercial momentum, while its valuation has risen substantially since January. Its next challenge will be turning that momentum into a durable position in a market where customers have multiple security products to evaluate and complex environments to protect.

The $400 million round gives the company resources to pursue that goal. Its longer-term progress will depend on whether runtime intelligence, contextual risk analysis, and automated security workflows translate into outcomes customers can consistently measure.

Subscribe now

Relieving Hydrostatic Pressure Naturally

One of the biggest threats to any municipal retaining wall is water. When a severe storm hits, rainwater soaks into the soil behind a solid concrete wall. With nowhere to go, the water builds immense pressure against the back of the concrete. Engineers call this hydrostatic pressure, and it is the primary reason traditional retaining walls crack, lean, or completely fail. Cities have to install complicated, expensive drainage pipe systems just to keep concrete walls standing.

Rock-filled wire structures naturally solve this problem. Because the baskets are filled with loose, angular stones, they feature thousands of tiny, built-in voids. When groundwater builds up behind the wall, it simply flows right through the rocks and out the front. The water never has a chance to build up dangerous pressure. This natural permeability makes these walls the perfect choice for lining flood channels, stabilizing riverbanks, and managing municipal stormwater runoff. They manage the flow of water safely without requiring secondary drainage systems that eventually clog and fail.

Adapting to Shifting Ground

The earth beneath our feet is rarely still. Soil expands when it freezes, shrinks during a drought, and constantly settles over time. When a city builds a long concrete retaining wall along a highway embankment, that rigid structure cannot adapt to the changing ground. Even a slight shift in the foundation will cause a solid wall to crack, requiring expensive and disruptive municipal repairs.

Wire and rock structures don't have this problem because they aren't rigid; they are inherently flexible. If the ground settles unevenly beneath a rock wall, the wire mesh simply bends, and the stones shift slightly to accommodate the new shape. The structure maintains its full integrity even as it contours to the changing landscape. This flexibility is vital for civil engineering projects built on soft soil or in areas prone to minor seismic activity. Municipalities can build long, structural embankments knowing the wall will bend and settle rather than snap under the stress.

Lowering Installation and Maintenance Costs

Every municipal project operates on a tight budget. City planners are constantly looking for ways to stretch taxpayer dollars without compromising safety or quality. Pouring concrete is incredibly expensive. It requires building complex wooden forms, routing specialized mixing trucks to the site, and hiring highly trained crews to manage the pour. Plus, concrete takes weeks to fully cure before it can bear any significant load.

Building with wire enclosures is a much simpler, faster process. A standard road crew can easily assemble the empty cages on-site and fill them using standard earth-moving equipment. There isn't any waiting around for cement to dry; the wall is at full structural strength the moment the last rock is placed. Cities can often source the fill stone from local quarries, cutting down on transportation costs. Over the lifespan of the structure, maintenance is virtually non-existent. You don't have to patch cracks or paint over graffiti, which frees up city maintenance crews to focus on other pressing infrastructure needs.

Blending Infrastructure with Nature

Beyond structural benefits, municipalities are also becoming more conscious of how their infrastructure looks. A towering, gray concrete wall feels cold and industrial. It creates an eyesore that residents often complain about, especially in parks, along greenways, or near residential neighborhoods.

Rock walls offer a much softer, organic aesthetic. The natural tones of the stone blend seamlessly into the surrounding landscape. More importantly, these structures actually encourage local ecology. Over the years, windblown dirt settles into the small crevices between the rocks. Seeds take root, and soon, native grasses, mosses, and vines begin weaving their way through the wire. What starts as a structural retaining wall eventually becomes a living, breathing part of the local ecosystem. For city planners trying to balance necessary infrastructure with environmental conservation, this ability to integrate into the landscape is a huge selling point. The wall supports the earth while providing small micro-habitats for local insects and plants.

Building for the Future

Managing a city's infrastructure is a tough balancing act. Municipal engineers need solutions that are structurally sound, cost-effective, and environmentally responsible. While solid concrete walls will always have a place in commercial construction, it's easy to see why so many local governments are shifting their focus to more organic methods. By embracing permeable, flexible rock structures, cities can safely manage storm runoff, stabilize tricky highway embankments, and create long-lasting infrastructure that doesn't drain the annual maintenance budget. It's a practical, rugged approach to civil engineering that proves sometimes the simplest materials really do provide the best long-term solutions for our communities.

Subscribe now

For cybersecurity teams, the arrival of new workplace technology often creates a familiar dilemma: restrict access and risk slowing employees down, or allow adoption and accept unfamiliar risks. Artificial intelligence is making that trade-off more complicated by enabling employees to build software, launch agents, and automate tasks with tools that were not traditionally part of their jobs.

The funding announcement for Pluto Security was first reported by TechCrunch. The company has now emerged from stealth with $42 million in total funding, multi-million-dollar annual recurring revenue, and more than 1 million endpoints secured globally. Its customer base includes more than 40 enterprises, among them Oscar Health, Workato, Sixt, AlphaSense, and Fortune 500 organizations.

From Blocking to Understanding

Led by CEO and co-founder Shahar Bahat, Pluto is building a security layer for the AI workspace. Its platform is designed to give enterprises visibility, contextual risk assessment, and real-time control across the AI tools employees use, the supply chain around those tools, and the applications, agents, and automations they create.

Bahat’s description of the company’s philosophy starts with a criticism of security models that rely primarily on stopping activity.

“Security has traditionally been built around stopping things, but that model breaks down when the business is moving faster than security can evaluate every new tool,” she said.

Pluto instead seeks to understand what employees are trying to accomplish and provide security teams with context to make more precise decisions. The idea is to enable legitimate work while restricting unsafe actions, rather than treating every unfamiliar tool as a threat that must be blocked.

That philosophy extends to the company’s engagement with security professionals. Bahat said Pluto listens to practitioners, researches the risks they encounter, and creates tools intended to help them move forward safely.

The Changing Definition of an Enterprise Workspace

AI is broadening who can create software inside an organization. Claude Code, Codex, Cowork, Lovable, and Base44 allow employees to develop applications, agents, automations, and workflows without depending entirely on traditional development teams.

However, the resulting environment involves more than the initial AI application. Skills, extensions, packages, MCPs, models, and other dependencies can become part of the AI supply chain. Pluto says employees across its customer base use hundreds of supply chain components while building thousands of applications.

The company’s platform combines endpoint visibility, contextual analysis, and real-time enforcement to help security teams manage this activity. Its capabilities include blocking malicious skills, restricting unsafe agent actions, and preventing access to sensitive data or credentials without stopping legitimate work.

This approach reflects Pluto’s view that security teams need to understand how AI tools are being used, not merely maintain lists of approved and prohibited software.

A Company Growing Through Research and Community

Pluto’s development has included a focus on resources that practitioners can use outside the commercial product. Its Plutonium hub, powered by proprietary RisksDB, helps security professionals assess risks across AI supply chain components such as skills and extensions. Pluto reports that thousands of security professionals are already using and contributing to the resource.

Saaya Pal, partner at Jump Capital, cited this model as a distinguishing feature of the company.

“The clearest example is Plutonium, their AI risk database — there was no good public data on what AI tools are actually doing inside enterprises, so the team built it and gave part of it away,” Pal said.

According to Pal, hundreds of practitioners use Plutonium every week, and most of the company’s pipeline now comes inbound. She said Jump Capital led Pluto’s latest round in part because distribution is difficult to retrofit later.

The approximately $34 million financing round brings Pluto’s total funding to $42 million. Its seed round was led by TLV Partners, with participation from additional investors.

An Experienced Team With a Broader Ambition

Bahat founded Pluto alongside Tal Ofer Regev and Gil Maman. The founding team has experience spanning Microsoft, Palo Alto Networks, Bridgecrew, and Israel’s Unit 8200.

Bahat previously held senior cybersecurity leadership roles at Microsoft and Palo Alto Networks and is a Forbes 30 Under 30 honoree. Regev previously served as vice president of research and development at Palo Alto Networks, where he managed more than 150 people. Maman spent more than six years in Israeli Military Intelligence in technical leadership roles spanning advanced cyber research and security engineering.

The company has grown to more than 50 employees across the United States and Israel. Its investors include Jump Capital, TLV Partners, Skinos, AlphaDrive Ventures, Ansa Capital, MTF, and individual backers with experience across cybersecurity and technology, including Ariel Maislos, Yevgeny Dibrov and Nadir Izrael of Armis, Slavik Markovich of Descope, Asaf Ezra of Andreessen Horowitz, and Sharin Fisher Dibrov and Idan Tendler of Vine Ventures.

Making Room for AI Without Losing Control

Pluto plans to use its new capital to invest in product development, research, and U.S. expansion as the AI supply chain evolves.

Rona Segev, co-founder and managing partner at TLV Partners, said the company’s approach gives security teams a way to understand what is happening and act with precision.

“As AI reshapes how people work, Pluto gives security teams a way to understand what’s happening and act with precision, rather than simply blocking innovation,” Segev said.

With reported multi-million-dollar ARR, more than 1 million endpoints secured, and over 40 enterprise customers, Pluto is entering its next stage with a product built around the tension between enabling AI adoption and managing its risks. Whether that model can keep pace with the changing AI workspace will be central to the company’s continued expansion.

Subscribe now

Instead of focusing purely on what their birth certificate says, the decision should be based on their behavioral track record, financial readiness, and actual driving experience.

Evaluating Maturity and Decision-Making

Does your teenager make smart choices under pressure? Driving is a highly demanding cognitive task that requires constant situational awareness and rapid decision-making. If your teen is naturally impulsive, frequently struggles to follow basic household rules, or easily caves to peer pressure, putting them in solo control of a two-ton machine is likely premature.

Establishing strict ground rules, like a zero-tolerance policy for distracted driving and setting passenger limits, is essential before letting a newly licensed driver go out alone. You need to know they’ll follow the rules when you’re not in the passenger seat. If you can’t trust your teenager to leave their smartphone alone during family dinner, you certainly can’t trust them to ignore a group text while navigating a busy four-lane highway.

The Financial Reality of Ownership

Vehicles remain expensive long after the initial purchase is finalized. Before you even begin browsing local inventory, sit down for a transparent conversation about the ongoing costs of fuel, insurance premiums, routine maintenance, and unexpected mechanical repairs. Will you be footing the entire bill, or are they expected to contribute a percentage of the costs?

Waiting until your teen has a part-time job or a consistent way to chip in for expenses is a highly effective strategy. When teenagers have a direct financial stake in their ride, they’re statistically more likely to drive cautiously and take care of basic maintenance like oil changes and tire pressure checks. Forcing them to pay for their own gas or a portion of the insurance premium transforms the vehicle from a free toy into a serious, tangible investment that requires respect.

Prioritizing Safety Over Style

When you ask a teenager what kind of car they want, they usually prioritize aesthetics, speed, or the stereo system. Parents need to ignore those requests completely and focus entirely on crashworthiness and safety ratings. When you decide the timing is right, avoid the temptation to put them in an oversized pickup truck or a tiny, fast sports car. Large trucks and older SUVs can be difficult for inexperienced drivers to maneuver and park, while fast cars inherently encourage speeding.

Instead, target a midsize sedan or a small crossover with a proven track record. Look for vital features like electronic stability control, strong roof strength, and high overall safety ratings. Reliability and structural safety must trump curb appeal every single time.

Experience Trumps Age

Getting a driver’s license simply means your teen passed a standardized test. It doesn’t make them a seasoned driver. Many families find that waiting six months to a full year after their teenager gets licensed before buying them a dedicated vehicle is the safest approach. During this probation period, the teen can simply share the primary family vehicle.

Sharing requires them to ask for permission, coordinate schedules with the rest of the household, and drive strictly under your specific conditions. It gives them critical time to build up their hours on the road in various weather conditions, from heavy rain to winter sleet, and to navigate complex traffic patterns before they are handed the ultimate freedom of their own set of keys.

The Right Time is a Family Decision

You know your teenager's capabilities better than anyone else. Watch how they handle their current, everyday responsibilities. Look at their schoolwork, their chores, their part-time jobs, and how well they respect curfews. If they consistently show up, follow through on their promises, and demonstrate solid judgment in their social lives, they’re likely ready for the serious responsibility of vehicle ownership. Conversely, don’t hesitate to delay the purchase if they’re still struggling with basic impulse control or pushing boundaries. A vehicle is a massive privilege, and making them wait until they are truly ready protects them, your finances, and everyone else sharing the road.

Take Over the Heavy Lifting

As her body changes, mundane chores quickly become exhausting. Don’t wait for her to ask for help, because many women feel guilty admitting they can’t keep up with their usual routines. Saying "let me know if you need anything" places the mental burden of delegating tasks onto her. Instead, make specific offers. Text her and say you are coming over on Saturday morning to scrub the bathrooms, vacuum the stairs, or walk her dog. If you know she’s having a tough week, show up with a bag of groceries and take care of the meal prep. Taking the physical labor off her plate gives her immediate permission to rest.

Be a Safe Space for Her Worries

We often treat pregnancy as a purely joyful experience, but it also brings sudden hormone shifts and physical discomfort. She might be worried about labor, the financial strain of childcare, or how her career will change. Be the person she can vent to without judgment. Resist the urge to offer unsolicited advice or tell her she’s overreacting. Sometimes, she just needs to complain about her swollen ankles. If she has specific medical concerns, encourage her to consult reliable resources rather than random internet forums.

Help Navigate the Appointment Overload

Expecting mothers have a dizzying number of appointments on their calendars. If her partner is working or unable to attend a checkup, offer to be her designated driver and note-taker. Doctor visits can be overwhelming, and having a second pair of ears to listen to the doctor's instructions can be a huge relief. You can also help her track her baby's development so you know what she is experiencing. Following a timeline allows you to understand exactly what physical symptoms she’s managing during each trimester, making you a much more empathetic support system.

Keep Her Fed and Hydrated

Food aversions and bizarre cravings are very real, especially during the first trimester when nausea is at its peak. The foods she normally loves might suddenly make her sick, and the only thing she can stomach might be plain bagels or fruit popsicles. Pay attention to what she can tolerate and keep her pantry stocked with those specific items. Drop off a high-quality insulated water bottle to encourage her to stay hydrated. As she gets closer to her due date, start bringing over freezer-friendly meals that she can easily heat up after the baby is born.

Remember Who She Is Outside of Motherhood

Once a woman announces she’s expecting, it’s common for people to stop asking about anything else. Every conversation suddenly revolves around the nursery theme or her due date. While she’s undoubtedly excited about the baby, she’s still the exact same sister you grew up with. Make a conscious effort to talk about other things. Gossip about your favorite television shows, complain about your mutual friends, or discuss the books you’re reading. Take her out for a nice dinner that has absolutely nothing to do with babies. Reminding her that her identity is not solely tied to motherhood is incredibly refreshing.

Stepping up for your sister right now sets the tone for your relationship once the baby finally arrives. Showing her that she can rely on you for the unglamorous tasks (like cleaning and cooking) proves you’re in her corner for the long haul. Being an incredible aunt or uncle starts long before the birth, and your consistent presence during these nine months will mean more to her than any physical gift ever could.

If you want to experience something that feels almost like magic, bioluminescence tours offer an adventure dictated entirely by the rhythms of nature. You get to glide through water that glows electric blue with every stroke of your paddle. But what exactly makes these nighttime excursions so radically different from your standard sightseeing trips?

An Interactive, Living Phenomenon

On a standard wildlife cruise, you sit back with binoculars and hope a dolphin or bird flies past. You’re mostly a passive observer. When you paddle out into a glowing waterway, you become part of the ecosystem's reaction.

The glowing effect is typically caused by millions of microscopic organisms called dinoflagellates. These tiny marine plankton emit light as a defense mechanism when the water around them is disturbed. According to the National Ocean Service, this chemical reaction produces light energy with almost no heat, a phenomenon known as cold light.

Because the light is only triggered by movement, you have to interact with the environment to see the show. Every time your kayak paddle slices through the surface, a swirling trail of blue or green light follows. If a school of mullet darts under your boat, they look like glowing underwater comets. Your movement is the thing creating the phenomenon around you.

Dictated by the Moon, Not a Clock

Traditional tourist attractions operate on fixed business hours. You can visit a museum on any given Tuesday at noon, rain or shine. Bioluminescent waters, however, are notoriously picky about when they want to show off.

First, the ambient light matters immensely. To see the glow clearly, you need absolute darkness. This means operators have to schedule trips around the lunar cycle, with the absolute best viewing occurring during a new moon. If you book a trip during a full moon, the brightness in the night sky will wash out the glow in the water. Planning around moon phases is something you rarely have to think about when booking a standard walking tour.

Second, temperature and water quality dictate the concentration of the plankton. Many hotspots only experience this glowing phenomenon during specific summer months when the water is warm enough for the microorganisms to multiply rapidly.

A Completely Different Sensory Focus

Most guided daytime trips overload you with information. A guide speaks into a microphone, pointing out architectural details or rattling off historical dates while you look around in broad daylight.

Paddling out in the pitch black shifts your senses completely. Because you can’t see the shoreline clearly or the people in the boats next to you, your hearing and touch become heightened. The primary sound is just the rhythmic splash of water and perhaps the distant call of a night bird. There are no loud city noises, passing traffic, or bright streetlamps to distract you. This creates a deeply meditative, quiet environment that traditional, bustling tourist traps simply can’t replicate. Guides on these trips tend to speak in hushed tones, allowing the serene, glowing water to hold your full attention.

Embracing the Unpredictable

When you buy a ticket to an amusement park, the roller coaster runs the exact same way every single time. Nature, on the other hand, makes no guarantees. The intensity of a glowing river or lagoon changes from night to night. A heavy afternoon thunderstorm might dilute the salinity of the water, temporarily dimming the plankton. A sudden temperature drop could cause the organisms to move deeper into the water column or shift to a different area entirely.

This unpredictability is exactly what makes the adventure so highly sought after. You’re taking a chance on a rare natural event. When the conditions align perfectly and the water lights up like a liquid starry sky, the payoff is incredibly rewarding precisely because it is fleeting and wild.

Leaving the Beaten Path

Ultimately, skipping the crowded daytime excursions for a midnight paddle is about changing how you interact with a destination. Standard tours show you a place as it has been built and controlled by people, while a dark, glowing waterway reveals how the earth behaves when we simply let it be. If you’re tired of the usual crowded beaches, predictable sightseeing itineraries, and bright neon signs, trading them for a quiet kayak and a paddle full of blue light might be exactly what your next vacation needs.

To keep your home comfortable and safe, you need to prepare your heating equipment before the freezing temperatures arrive. Scheduling professional HVAC service is the best way to ensure everything runs smoothly, but you can also tackle several tasks yourself. Taking a proactive approach to your fall furnace maintenance will save you money on utility bills and prevent inconvenient breakdowns. Let's explore the essential steps you should take this autumn to get your system ready for the cold weather.

Replace or Clean the Air Filter

The simplest and most effective maintenance task you can perform is changing your air filter. Over the summer, your cooling system pushed air through this filter, trapping dust, pet hair, and pollen. If you leave a clogged filter in place when you switch to heating, your furnace has to work incredibly hard to push warm air through that barrier. This restricted airflow forces the blower motor to run longer, which drives up your energy bills and causes unnecessary wear and tear on the equipment.

Locate your filter and pull it out to inspect its condition. If you can't see light through it, it's time for a replacement. Most standard filters need to be swapped out every one to three months, depending on whether you have pets or suffer from allergies. Write the date on the new filter's cardboard frame before sliding it in so you know exactly when you changed it. Starting the season with a fresh filter ensures optimal airflow and keeps your indoor air quality high.

Clear the Area Around the Furnace

Basements, garages, and utility closets often turn into storage areas during the spring and summer. It's common to find boxes of holiday decorations, old furniture, or laundry baskets stacked right next to the heating unit. However, your furnace needs adequate space to operate safely and efficiently.

Gas systems require a steady supply of oxygen to burn fuel properly. If you block the surrounding area with clutter, you starve the unit of combustion air. Storing cardboard, clothing, or cleaning chemicals near a heat source creates a serious fire hazard. Take an afternoon to declutter the utility room. You should maintain at least three feet of space on all sides of the equipment. This clearance keeps your home safe and gives a technician enough room to work when they arrive for an inspection.

Test the Thermostat

Your thermostat acts as the brain of your heating system. If it isn't functioning correctly, your home won't reach the desired temperature, no matter how healthy the furnace is. Before the first real cold snap hits, switch the device from cool to heat and turn the temperature up a few degrees higher than the current room temperature.

Listen closely. You should hear the system click and feel warm air blowing from the vents within a minute or two. If the system doesn't respond or shuts off immediately, you might have a problem. Often, a malfunctioning thermostat just needs fresh batteries. Swap out the old batteries for new ones and try the test again. If you have a smart thermostat, check the app to ensure your heating schedule is set up correctly for the season and that the software is up to date.

Clean and Unblock Your Vents

Warm air can't circulate if it doesn't have a clear path into your rooms. Walk through your house and inspect every supply vent and return grille. Over the summer, dust and pet dander likely settled inside the grilles. Use a vacuum attachment to clean out this debris so it doesn't blow all over your living space the first time the heater kicks on.

While you do this, check to make sure furniture, rugs, and thick curtains aren't blocking the airflow. It's tempting to place a sofa right over a floor register to make a room look better, but doing so traps the heat and throws your entire system off balance. Keeping all vents open and unobstructed allows the system to heat your home evenly and efficiently.

Inspect the Visible Exhaust Flue

For homes with gas heating, the exhaust flue carries dangerous combustion gases safely outside. Take a quick walk around the exterior of your home and locate the exhaust pipe. Depending on your system's efficiency rating, this might be a metal pipe on your roof or a white plastic pipe sticking out of the side of your house.

Check the opening to ensure it isn't blocked by leaves, overgrown bushes, or bird nests. An obstructed exhaust pipe forces harmful gases to back up into your home, which triggers safety switches that shut the system down. While you're looking at the pipe, verify that there aren't any visible cracks, rust spots, or disconnected joints. If you spot any damage, don't try to patch it yourself; call a professional to handle the repair.

Check Your Carbon Monoxide Detectors

A gas furnace burns fuel to create heat, and carbon monoxide is a natural byproduct of that combustion process. Under normal conditions, the system vents this dangerous gas safely outside. However, if the heat exchanger develops a crack or the exhaust pipe gets blocked, carbon monoxide can leak into your living space. Since this gas is invisible and odorless, you rely entirely on your detectors to keep your family safe.

Fall is the perfect time to test every detector in your house. Press the test button to ensure the alarm sounds loudly. If the device chirps or fails to sound, replace the batteries immediately. Keep in mind that carbon monoxide detectors expire after five to seven years. Check the manufacture date on the back of the unit; if it's expired, replace the entire device to guarantee it functions when you need it most.

Schedule a Professional Tune-Up

While you can handle filters and vent cleaning on your own, the internal components of your heater require a trained eye. You shouldn't try to adjust gas pressure, clean delicate flame sensors, or inspect electrical connections without professional training.

A licensed technician has the right tools to perform a comprehensive safety and performance inspection. They'll clean the burners, test the safety switches, tighten loose wiring, and inspect the heat exchanger for microscopic cracks. Catching a failing part in October is much cheaper and easier than dealing with a complete breakdown in December. Regular professional maintenance extends the lifespan of your equipment and gives you total peace of mind as you head into the coldest months of the year.

Perform Daily Visual Inspections

The easiest way to prevent a catastrophic tear is to spot the damage early. Before hitting the road, walk around your trailer and inspect the fabric. Look closely where the material rests against the metal bows and end caps. These contact points experience the most wear as the truck bounces down the highway. If you spot any tiny pinholes or frayed edges, repair them immediately with a high-quality patch kit.

You shouldn't just look at the fabric. Inspect the bungee cords, tie-downs, and straps. If a strap looks worn or a bungee has lost its elasticity, swap it out. A loose strap allows the wind to get underneath, causing the material to flap rapidly. That constant whipping motion degrades the reinforced seams quickly. A quick visual check saves you from dealing with a shredded cover halfway through your route.

Clean the Tracks and Hardware

Trailers operate in dusty, dirty environments. Grain dust, gravel particles, and road salt naturally accumulate inside aluminum tracks and around moving hardware. When this debris builds up, it forces the mechanical parts to work harder to roll the cover. Over time, that extra resistance burns out electric motors and bends manual crank arms.

Make it a habit to clean out the tracks and channels at least once a week. Use a stiff-bristled brush or an air compressor to blow dirt from the tracks, pulleys, and springs. Pay special attention to the front and rear caps, where debris tends to pool. Keeping the hardware clean ensures the system rolls smoothly and prevents abrasive grit from wearing down your expensive aluminum components.

Lubricate Moving Components

Moving metal parts require proper lubrication to function correctly, but you have to use the right type. Standard oil-based sprays or thick greases are usually a bad idea for trailer hardware. Grease attracts dust, creating a grinding paste that ruins bearings and tracks.

Instead, use a dry silicone or graphite spray on your mechanical joints. Spray the pivot points, the universal joints on your crank arm, and any exposed gears. If you operate an automated setup, check the manufacturer's recommendations for lubricating the motor's drive components. Applying a dry lubricant every few weeks keeps everything moving freely without turning your hardware into a magnet for road grime.

Manage Your Tarp Tension

Finding the right tension is crucial for maximizing the life of your cover. If the fabric is too loose, the wind catches it. A flapping cover acts like a parachute, pulling against the bows and stressing the anchor points until the material tears. If the cover is pulled too tight, it puts unnecessary strain on the roller tube, the motor, and the seams.

You want a snug, secure fit that prevents wind from getting underneath, but it shouldn't look like a stretched drum. Weather conditions also impact tension. In the summer, heat causes the material to expand slightly, requiring you to tighten the straps. In the winter, freezing temperatures cause the fabric to shrink and stiffen. Adjust your tension settings seasonally to ensure the material stays secure without placing undue stress on the hardware.

Inspect the Electrical System

For automated setups, the electrical components require just as much attention as the physical fabric. The wiring running along your trailer is constantly exposed to the elements. Inspect the entire wiring harness for cracked insulation, exposed copper, or loose connections. Road salt and moisture quickly corrode bare wires, leading to short circuits.

Check the connections at the battery and the switch to ensure they're tight and free of rust. If you notice powdery corrosion on the terminals, clean it off with a wire brush and apply a terminal protector spray. Pay attention to the motor itself. If it sounds like it's struggling or rolls slower than usual, don't ignore it. A sluggish motor usually indicates a voltage drop or a binding track. Diagnosing the issue early prevents the motor from burning out entirely.

Store Your Trailer Smartly

When the busy season ends and your trailer sits idle for an extended period, how you store it makes a big difference. Leaving the trailer parked under a tree exposes the cover to acidic bird droppings, falling branches, and tree sap, all of which degrade the fabric's protective coating.

Whenever possible, park your equipment under a shed or inside a garage during the off-season. If indoor storage isn't an option, make sure the cover is fully closed and securely fastened. Standing water pooled in the center of the fabric stretches the material and strains the supporting bows. Ensure your trailer is parked on level ground so rain and melting snow run off the sides properly.

Taking care of your equipment isn't complicated, but it requires consistency. By keeping the hardware clean, managing your tension, and patching small holes early, you reduce your repair bills significantly. A little bit of proactive maintenance guarantees your cover will protect your cargo reliably, season after season.

Subscribe now

The enterprise AI market is entering a phase where implementation matters as much as experimentation. Companies are looking beyond model capabilities toward technologies that can connect AI to existing systems, manage its use, interpret information and turn it into practical business workflows.

The San Francisco Tribune reveals list of top startups working across those areas ahead of HumanX Amsterdam, which runs September 22–24 at RAI Amsterdam. The event will bring together more than 2,500 attendees, with startups, investors, enterprise buyers and technology leaders represented and 60% of attendees at VP level or above.

Making AI More Dependable

For many organizations, the challenge is not simply adopting AI but making it dependable enough for real operations.

causaLens is focused directly on this issue through Digital Knowledge Workers built with multi-agent architectures. Its platform includes pre-built Blueprints, a Factory for customized workers and the System of Work governance layer. The company emphasizes causal reasoning, human-in-the-loop controls, automated evaluation, monitoring and self-healing.

Trendium approaches enterprise adoption from the governance side. Its ContextGuard platform is designed to provide visibility and control over AI infrastructure, while its AI Enablement Program helps engineering teams adopt AI coding agents responsibly. Its framework is organized around visibility, control, enablement and compliance.

Connecting AI to Existing Information

Another challenge is giving AI access to information that businesses already depend on.

Unfold is developing technology for closed, undocumented and vendor-locked systems. The company says its approach can make these systems readable without an API, documentation or vendor involvement. Operating within the customer environment, it uses read-only access by default and provides full auditing.

Kensho Technologies takes a data-focused approach. Operating as S&P Global’s innovation engine, it combines AI and engineering with financial and business information. Its technology connects AI applications, agents and large language models to trusted S&P Global data while structuring proprietary information for machine learning and generative AI workflows.

From Speech to Specialized Discovery

The AI opportunity is also expanding into specialized forms of information processing.

AssemblyAI provides infrastructure for applications that rely on spoken language, with APIs supporting speech-to-text, contextual understanding and real-time agentic workflows. Speechmatics focuses on multilingual speech recognition, supporting more than 55 languages and multiple deployment environments. Its applications include healthcare, legal transcription, contact centers, meeting platforms, live captioning and AI voice agents.

Otter AI is approaching workplace conversations as a source of organizational knowledge. Its platform captures and transcribes meetings, identifies decisions and action items, and allows users to search and query accumulated meeting information. It also supports follow-ups and connections with workflows such as CRM systems.

Nuritas is applying AI to an entirely different information challenge: identifying useful peptides from nature. Through its Nuritas Magnifier platform, the company says it has identified more than 8 million peptides and built a library of peptides with known functionalities, with the goal of taking discoveries from computational prediction through clinical validation.

AI Meets Visual Commerce

PhotoRoom represents another direction for applied AI, using its technology to support visual production for e-commerce. The platform serves individual sellers as well as large enterprises managing millions of product images, offering batch editing, automated quality assurance, brand controls and integrations with PIM, DAM and other commerce systems.

The common thread across these companies is not a single AI application. It is the infrastructure surrounding AI: how systems access information, how organizations control deployments, how conversations become usable knowledge and how AI-generated outputs fit into existing workflows.

The Practical Side of AI Adoption

HumanX Amsterdam offers a snapshot of that broader transition. The companies highlighted for the event are working across different industries and technical layers, but their focus increasingly sits on the practical conditions required for AI to deliver value.

As enterprise adoption expands, those conditions are becoming an increasingly important part of the technology conversation.

Subscribe now

Speeding Up Data Processing and Pattern Recognition 

What exactly does artificial intelligence do for metrology? Its biggest strength is processing speed and pattern recognition. When an optical scanner captures millions of data points on a factory floor, traditional software might take several minutes to process that information. Modern algorithms can crunch that point cloud almost instantly, comparing the scanned part to the original computer-aided design model. They easily spot trends over time, like predicting when a cutting tool is starting to wear down because the parts are slowly drifting out of tolerance. This predictive capability saves companies money and reduces scrap material. AI also helps filter out environmental noise from scans, smoothing out the data so inspectors can see a clearer picture of the manufactured part. It handles the tedious, repetitive tasks of data analysis, freeing up quality control professionals to focus on solving complex engineering problems rather than staring at spreadsheets all day.

The Dangers of Relying on Predictions Over Reality 

Despite these impressive capabilities, we have to remember what artificial intelligence actually is. It's essentially an advanced prediction engine. It doesn't physically touch the object, and it doesn't experience the real world. If a camera lens has a smudge or if the lighting in the factory changes, a vision-based AI system might misinterpret a shadow as a defect. Or worse, it might ignore a real defect because the lighting hid it. Software relies entirely on the raw data collected by sensors and probes. If that initial data isn't perfectly accurate, the algorithm's output will be flawed. You can have the smartest network available, but if the physical sensor isn't calibrated correctly, the system will confidently give you the wrong answer. You can't code your way out of poor physical measurement. The ground truth always comes back to the physical interaction between a calibrated probe and the surface of the part. If you trust a digital estimation over a physical measurement, you are playing fast and loose with your quality control standards.

Why Guessing Fails in High-Stakes Manufacturing 

One of the most concerning aspects of leaning too hard on algorithms for measurement is how they handle missing information. When a non-contact scanner misses a spot on a shiny or complex surface, the software often steps in to fill the gap. It looks at the surrounding data points and essentially guesses what that missing surface looks like. For a basic visual model, this interpolation works fine, but in precision engineering, a guess is a serious liability. If you manufacture parts for airplanes or medical implants, a fraction of a millimeter means the difference between success and catastrophic failure. You can't rely on a computer guessing the depth of a bore hole or the exact angle of a valve seat. Physical measurement tools touch the actual surface, providing concrete, undeniable data rather than a statistically probable estimation. In high-stakes manufacturing, certainty will always trump probability. You need to know without a shadow of a doubt that the component meets the engineering specifications, and the only way to achieve that level of confidence is through direct, physical verification.

Navigating Complex Geometries and Hidden Features 

Another area where artificial intelligence falls short is physical accessibility. Cameras and lasers operate on a line of sight. If a sensor can't see a feature, the software can't analyze it. Modern manufactured parts often feature incredibly complex internal geometries, deep intersecting holes, and hidden undercuts. You simply can't measure what you can't see. Tactile probes physically reach into these tight spaces, mapping out the internal dimensions with physical certainty. While software can help stitch together different scans or optimize the path the probe takes, it can't physically maneuver into a dark cavity to verify a tight tolerance. The physical world contains obstacles that lines of code simply can't navigate without the help of precision mechanics and physical hardware.

Managing Environmental Variables in Real Time 

Real-world measurement also involves managing physical variables like thermal expansion and material deflection. An algorithm doesn't feel the temperature drop on a factory floor, which might cause a metal component to shrink by a few microns. Physical coordinate measuring tools account for these physical realities, often using temperature sensors and specialized materials to maintain accuracy regardless of the environment. Software can apply compensation formulas, but it still relies on accurate real-time physical inputs to make those adjustments. If the initial physical data is skewed by temperature changes and the sensor doesn't catch it, the AI will process that flawed data as if it were perfect, leading to accepted parts that should have been rejected.

Embracing the Best of Both Worlds 

We are entering a fascinating era for manufacturing and quality assurance. Artificial intelligence brings undeniable benefits to the table, making data analysis faster, smarter, and more predictive. It helps teams spot trends before they become expensive problems and streamlines the entire inspection workflow, but we shouldn't confuse data processing with data creation. The foundation of metrology is, and always will be, rooted in the physical world. Taking a proper measurement requires calibrated hardware, environmental control, and direct interaction with the part being inspected. Algorithms make excellent assistants, but they make terrible substitutes for physical reality. By combining the processing power of modern software with the undeniable accuracy of physical measurement tools, manufacturers achieve the best of both worlds. They get the speed of the future without sacrificing the concrete reliability that keeps our world running safely.

The early morning hours, roughly between 6:00 AM and 9:00 AM, belong almost entirely to text messaging. Very few people want their phone ringing while they're drinking their first cup of coffee or stumbling through their morning routine. An incoming phone call early in the morning usually signals an emergency, which is why most of us instinctively avoid dialing anyone before standard business hours begin. Texting during this early window, however, is widespread. It acts as a gentle, low-pressure way to say good morning, coordinate carpools, or double-check the day's schedule. Because text messages sit quietly in a notification shade until the recipient's ready to read them, they don't interrupt morning rituals. As the clock approaches 10:00 AM, the shift toward voice calls begins. By mid-morning, people have arrived at work or settled into their daily tasks. Between 10:00 AM and 11:30 AM, phone call volume hits its first notable surge of the day. This is the prime window for business calls, doctor's appointments, and important check-ins that require immediate answers. People have cleared their initial morning inbox and are ready to tackle active conversations.

Around noon, phone habits shift again. Lunch hours, typically spanning 12:00 PM to 1:30 PM, bring a notable spike in personal text messages. Employees stepping away from their desks use this break to catch up on group chats, respond to family members, or scroll through messages they missed during morning meetings. Phone calls drop off slightly during lunch, as few people want to spend their break navigating a noisy environment or talking over a meal. The afternoon push from 2:00 PM to 5:00 PM brings an interesting split in behavior. Work-related phone calls remain high until around 4:00 PM, as people try to close out deals, resolve pressing issues, or finish project updates before the workday ends. However, as 5:00 PM approaches, personal calls begin to pick up. The evening commute, between 4:30 PM and 6:00 PM, is one of the most popular times for casual phone conversations. Drivers trapped in traffic or commuters riding the train often use hands-free devices to catch up with parents, spouses, or friends to unwind after a long day.

Once 6:00 PM rolls around, work-related calls vanish for most people, replaced by a surge in personal communication that lasts until roughly 9:00 PM. This evening block represents the highest volume of text messaging out of the entire 24-hour cycle. During these hours, people are relaxing at home, preparing dinner, or watching television. Texting thrives here because it seamlessly integrates with leisure activities. You can exchange dozens of texts with friends about weekend plans while simultaneously cooking or watching a movie. Family group chats come alive during the evening, filling up with photos, jokes, and updates from the day. At the same time, the hours between 7:00 PM and 8:30 PM are favored for longer, meaningful phone calls with family members or close friends who live far away. Unlike quick daytime business calls, these evening phone conversations tend to last much longer, as both parties finally have the free time to sit down and talk without checking the clock.

After 9:00 PM, phone activity begins a steady decline. Calling someone after 9:00 PM isn't considered polite unless you've specifically agreed on a late-night chat ahead of time. A ringing phone late at night creates unnecessary anxiety, as people immediately assume something's gone wrong. Texting continues late into the night, particularly among younger demographics, but even text volume tapers off by midnight. Silent modes, Do Not Disturb settings, and sleep schedules naturally bring mobile activity to a near standstill until the sun comes back up.

Ultimately, the choice between placing a call and sending a text comes down to urgency and emotional depth. Text messaging dominates the overall volume of daily phone use simply because it's asynchronous—it allows people to communicate on their own schedule without demanding immediate attention. That's why texting stays steady throughout the entire day, peaking during leisure hours. Voice calls, on the other hand, are reserved for specific windows where real-time clarity matters most. We call during mid-morning business hours to get work done, and we call in the early evening to build personal connections. Understanding these natural daily waves makes it much easier to reach people when they're actually ready to talk.

The early morning hours, roughly between 6:00 AM and 9:00 AM, belong almost entirely to text messaging. Very few people want their phone ringing while they're drinking their first cup of coffee or stumbling through their morning routine. An incoming phone call early in the morning usually signals an emergency, which is why most of us instinctively avoid dialing anyone before standard business hours begin. Texting during this early window, however, is widespread. It acts as a gentle, low-pressure way to say good morning, coordinate carpools, or double-check the day's schedule. Because text messages sit quietly in a notification shade until the recipient's ready to read them, they don't interrupt morning rituals. As the clock approaches 10:00 AM, the shift toward voice calls begins. By mid-morning, people have arrived at work or settled into their daily tasks. Between 10:00 AM and 11:30 AM, phone call volume hits its first notable surge of the day. This is the prime window for business calls, doctor's appointments, and important check-ins that require immediate answers. People have cleared their initial morning inbox and are ready to tackle active conversations.

Around noon, phone habits shift again. Lunch hours, typically spanning 12:00 PM to 1:30 PM, bring a notable spike in personal text messages. Employees stepping away from their desks use this break to catch up on group chats, respond to family members, or scroll through messages they missed during morning meetings. Phone calls drop off slightly during lunch, as few people want to spend their break navigating a noisy environment or talking over a meal. The afternoon push from 2:00 PM to 5:00 PM brings an interesting split in behavior. Work-related phone calls remain high until around 4:00 PM, as people try to close out deals, resolve pressing issues, or finish project updates before the workday ends. However, as 5:00 PM approaches, personal calls begin to pick up. The evening commute, between 4:30 PM and 6:00 PM, is one of the most popular times for casual phone conversations. Drivers trapped in traffic or commuters riding the train often use hands-free devices to catch up with parents, spouses, or friends to unwind after a long day.

Once 6:00 PM rolls around, work-related calls vanish for most people, replaced by a surge in personal communication that lasts until roughly 9:00 PM. This evening block represents the highest volume of text messaging out of the entire 24-hour cycle. During these hours, people are relaxing at home, preparing dinner, or watching television. Texting thrives here because it seamlessly integrates with leisure activities. You can exchange dozens of texts with friends about weekend plans while simultaneously cooking or watching a movie. Family group chats come alive during the evening, filling up with photos, jokes, and updates from the day. At the same time, the hours between 7:00 PM and 8:30 PM are favored for longer, meaningful phone calls with family members or close friends who live far away. Unlike quick daytime business calls, these evening phone conversations tend to last much longer, as both parties finally have the free time to sit down and talk without checking the clock.

After 9:00 PM, phone activity begins a steady decline. Calling someone after 9:00 PM isn't considered polite unless you've specifically agreed on a late-night chat ahead of time. A ringing phone late at night creates unnecessary anxiety, as people immediately assume something's gone wrong. Texting continues late into the night, particularly among younger demographics, but even text volume tapers off by midnight. Silent modes, Do Not Disturb settings, and sleep schedules naturally bring mobile activity to a near standstill until the sun comes back up.

Ultimately, the choice between placing a call and sending a text comes down to urgency and emotional depth. Text messaging dominates the overall volume of daily phone use simply because it's asynchronous—it allows people to communicate on their own schedule without demanding immediate attention. That's why texting stays steady throughout the entire day, peaking during leisure hours. Voice calls, on the other hand, are reserved for specific windows where real-time clarity matters most. We call during mid-morning business hours to get work done, and we call in the early evening to build personal connections. Understanding these natural daily waves makes it much easier to reach people when they're actually ready to talk.

Subscribe now

For years, cybersecurity leaders have faced a frustrating equation: the number of systems that need to be protected keeps growing, while the number of people available to protect them does not. Artificial intelligence could change that equation by giving security teams access to capabilities that previously required armies of specialized analysts.

OpenAI is now calling on the industry to move quickly in that direction. Its new open letter, “A Call to Action on Cyber Defense,” has attracted first-day support from Palo Alto Networks, Mate Security, and ServiceNow, placing the three companies among a select group of cybersecurity market leaders backing the initiative.

The timing is significant. As AI models become more capable, cybersecurity is becoming a race not simply to develop better detection technology, but to determine which side can deploy intelligent systems at greater scale.

The Cost Of Staying Secure

Security teams have never operated in a perfectly maintained environment. Organizations routinely carry outdated systems, unpatched vulnerabilities, excessive privileges and years of accumulated technical debt.

Addressing those problems manually can be expensive and slow. OpenAI argues that the arrival of increasingly capable AI creates an opportunity to change the economics of that work.

“We have a limited window to strengthen cyber defenses,” the letter states.

The company argues that cyber-capable AI can make security tasks faster and more accessible, potentially allowing smaller teams to cover significantly larger environments.

That matters particularly for critical infrastructure, where security budgets and staffing levels can be constrained even though the consequences of an incident can be severe.

From More Tools To More Capacity

OpenAI’s proposal is notable because it focuses on defensive capacity rather than simply adding another security product to an already crowded market.

The company calls on cybersecurity vendors to strengthen existing tools with AI, continuously test defenses against advanced cyber capabilities and provide hands-on assistance to organizations that may struggle to deploy sophisticated technology on their own.

The objective is essentially to make the work of a security team scale.

If AI can investigate vulnerabilities, analyze threats, validate remediation and support incident response, organizations could potentially increase their defensive coverage without increasing headcount at the same rate.

That is one reason the emergence of agentic cybersecurity companies is attracting attention.

Mate Security’s Different Proposition

Mate Security is approaching the problem by building an open foundation for agentic cyber defense.

Rather than focusing exclusively on AI-assisted analysis, Mate is developing systems in which AI agents can participate directly in investigations. Its Gamebooks technology provides structured investigation procedures while allowing agents to reason, pivot and act as an investigation unfolds.

The distinction is important from an economic perspective. An AI system that simply summarizes an alert may save an analyst minutes. An agent capable of conducting substantial portions of an investigation could potentially change how an entire security operation is staffed and organized.

That does not eliminate the need for human expertise. Instead, it potentially allows security professionals to spend more time on the highest-value decisions while AI handles larger amounts of repetitive or investigative work.

The Established Players Are Moving Too

The participation of Palo Alto Networks and ServiceNow shows that this shift is not confined to startups.

Both are established market leaders operating the large-scale platforms that organizations already rely on, and their support for the initiative underscores the extent to which AI is becoming integrated into the broader enterprise market.

They also address different parts of the capacity problem. Palo Alto Networks operates across the security controls and telemetry that enterprises depend on for detection and enforcement. ServiceNow operates in the enterprise workflows where remediation is assigned, tracked and verified, which is often where the cost of security work accumulates.

That second point matters to the economics. Faster investigation produces limited savings if the resulting fixes still move at the pace of manual coordination, and OpenAI’s letter asks organizations to measure progress by whether attacks are contained quickly and whether fixes actually work.

Mate Security represents a newer architectural direction, while Palo Alto Networks and ServiceNow illustrate the ability of established platforms to bring AI capabilities into existing enterprise environments.

Together, their participation points toward a market in which AI may increasingly become a baseline component of cyber defense.

The Race For Scale

OpenAI’s letter ultimately argues that defenders need to act before AI-enabled attacks become more sophisticated.

“Sharing tools, practical knowledge, and verified fixes lets one organization’s work help protect many others,” the company says.

That principle could become particularly important as organizations search for ways to improve security without simply increasing spending indefinitely.

The cybersecurity industry’s AI race may therefore be measured in a different way than previous technology cycles. The winners may not necessarily be the companies with the most advanced AI demonstrations, but those that can turn AI into reliable defensive capacity at the largest possible scale.

By joining OpenAI’s initiative as first-day signatories, Palo Alto Networks, Mate Security, and ServiceNow are signaling that the industry is moving toward that model. The question now is how quickly organizations can translate increasingly capable AI into measurable reductions in cyber risk.

Subscribe now

Subscribe now

Asset management has spent years absorbing machine intelligence into parts of the investment process. FINQ took a different route by handing the whole process over. The results reported for the period ending August 31, 2026 give that decision a track record to point at: 23.51% since inception for AIUP, 23.83% for AINT, and 11.61% for the S&P 500 across the same span beginning February 5, 2026.

Full Delegation Rather Than Partial Assistance

The distinction FINQ draws is structural. Its proprietary AI framework is a systematic approach aimed at beating the indexes by autonomously ranking, selecting, and weighting all participants of an index. Ranking, selection, and weighting are the three decisions that define a portfolio, and in this design all three are handled by the system.

That approach carries a formal marker. AIUP and AINT are the first SEC-registered ETFs in the United States to be fully managed by artificial intelligence. Both trade on NYSE Arca.

The Products in Practice

AIUP is FINQ’s long-only U.S. large-cap equity ETF. It has continued to outperform the benchmark while maintaining broad exposure to its top-ranked companies. As of August 31, 2026, AIUP shows a since-inception return of 23.51%, a net asset value of $29.97, and a market price of $29.99.

AINT is FINQ’s dollar-neutral strategy. It has been buying the top-ranked and selling short the lowest-ranked in its relative ranking AI model. Its since-inception return stands at 23.83%, with a net asset value of $30.97 and a market price of $30.96.

Placing a long-only and a dollar-neutral product on the same ranking engine tests that engine from two directions at once. AIUP relies on the top of the ranking being right. AINT relies on the ordering being right across its full range, since it takes positions at both ends.

A Record Measured Month by Month

FINQ reported not only the cumulative outcome but the shape of it. AIUP has outperformed the S&P 500 at every month-end since inception. AINT has outperformed the S&P 500 in every month except its first month of trading, and following that initial launch month it has consistently outperformed the S&P 500. The company presents AINT’s pattern as further demonstrating the adaptability of its AI framework.

For an industry that evaluates strategies on repeatability, the month-end series is the more informative half of the disclosure. It converts a single cumulative figure into a sequence of observations across seven monthly checkpoints, from the end of February through the end of August.

The two products carry that record under the full names FINQ FIRST U.S. Large Cap AI-Managed U.S Equity ETF and FINQ Dollar Neutral U.S. Large Cap AI-Managed U.S Equity ETF, both launched on the same day on the same exchange.

Speed as a Design Argument

FINQ, an AI-driven asset management company focused on building autonomous investment products, reports that its proprietary AI system systematically evaluates vast amounts of financial and market data of each index participant in real time, enabling the ETFs to dynamically adjust holdings based on evolving market conditions. The company holds that the sustained outperformance highlights the framework’s ability to consistently identify and capitalize on market opportunities with speed and precision that traditional human-managed models cannot replicate.

That is a claim about throughput as much as insight. A framework reading every participant in an index continuously operates at a cadence that differs in kind from periodic committee review.

The View From the Top of the Company

“These results demonstrate the strength and consistency of our AI framework during dynamic market environments,” said Eldad Tamir, founder and CEO of FINQ. “I believe autonomous investing will continue to reshape asset management, and the performance of AIUP and AINT reflects the growing ability of AI to adapt, identify opportunities, and respond to market changes at scale.”

FINQ believes it can continue to build on its early gains following its February launch on NYSE Arca.

Subscribe now

Subscribe now

Every dollar a digital business earns passes through its network edge. Checkout requests, API calls, session tokens, mobile app traffic, partner integrations: all of it crosses the same boundary before it reaches an application. That boundary has traditionally been budgeted as a security cost. Huskeys is arguing it should be understood as revenue infrastructure.

The Wall Street Journal first reported that the company has closed a $27 million Series A led by Blackstone Innovations Investments, bringing total funding to $35 million after an $8 million seed round.

The Commercial Logic of Getting the Edge Right

Huskeys is building a Network Edge Security Management platform, a category it abbreviates as NESM. The commercial case for that platform rests on a simple observation: decisions made at the edge have direct financial consequences in both directions.

Serving legitimate traffic accurately keeps customers transacting and revenue flowing. Filtering malicious traffic accurately keeps the organization protected. The requirement is to strengthen security while minimizing false positives, maintaining business continuity, and protecting revenue-critical applications. Those goals are usually presented as a trade-off. Huskeys’ position is that better management makes them compatible.

Itai Gafni, CEO and Co-Founder of Huskeys, put the point in commercial terms: “Every business today runs through its network edge - that’s where your customers, your revenue, and your threats all meet, yet for most organizations, that edge works against them instead of for them.”

Why the Financial Case Is Getting Stronger

Two developments are raising the stakes on edge decisions.

The first is the volume of automated traffic that carries commercial value. Autonomous agents are becoming legitimate users of applications, and non-human traffic is expected to account for 70 percent of all web traffic by 2027. As agents book, purchase, compare, and integrate on behalf of customers, a policy that treats automation cautiously by default starts applying to the majority of inbound commerce.

The second is the pace of the adversarial side. Attackers are using AI to discover and exploit vulnerabilities faster than ever before, and developments such as Mythos demonstrate how AI-driven capabilities are accelerating the speed and scale of cyber threats. Faster discovery compresses the time an organization has to respond after a vulnerability becomes public.

Time to Mitigation as a Financial Metric

That compression is why Virtual Patching reads as a business capability rather than a purely technical one. It allows security teams to mitigate vulnerabilities directly at the network edge within minutes, providing protection while a permanent fix is being developed, tested, and deployed in the application code.

The value is in the delta. Application code fixes move at the speed of an engineering release process. Edge mitigation moves at the speed of a policy push. Shortening the exposure window from a release cycle to minutes reduces risk without pulling engineers off roadmap work to expedite an emergency deployment.

Continuous posture assessment, dynamic policy generation, and orchestration round out the capability set, together helping organizations respond to emerging threats faster.

The Efficiency Argument

There is also a straightforward operating expense dimension. Modern organizations secure their applications across an expanding mix of cloud providers, Content Delivery Networks, Web Application Firewalls, and other edge services. Every one of those tools requires configuration, monitoring, and staff who understand it.

Huskeys’ platform operates on top of existing network infrastructure. Rather than replacing existing investments, it connects and orchestrates the modern network edge ecosystem, including CDNs and WAFs, cloud-native security solutions, and network components such as load balancers, VPCs, and security groups.

That framing protects prior spending. A buyer does not have to write off a multi-year CDN commitment or retrain a team on a new WAF to adopt the platform. The purchase is additive, which shortens the internal approval path considerably compared with a replacement proposal.

The patented Unified Data Model is what makes the additive approach work at a technical level. It creates a common layer of understanding across different environments, technologies, and providers, allowing security insights, policies, and actions to move seamlessly across platforms. The model supports multi-cloud and multi-vendor environments and helps organizations maintain consistent security from the Internet Edge all the way to the application.

Who Wrote the Checks

Blackstone Innovations Investments led. The round also included Merlin Ventures, Skinos Ventures, Zscaler Ventures, Okta Ventures, Bright Pixel Capital, and SV Angel. Skinos Ventures was founded by Shlomo Kramer and Yishay Yovel, pioneers of the WAF and SASE categories.

Individual investors include Eran Reshef, inventor of the WAF and the CAPTCHA, along with executives from Palo Alto Networks, Cloudflare, Check Point, AWS, Google, Microsoft, and Intel.

For a company selling on commercial outcomes, the presence of Zscaler Ventures and Okta Ventures is worth noting. Both are strategic arms of businesses that sell into the same buyers. Their participation indicates the platform is read as an expansion of the edge security budget rather than a reallocation of it.

Adam Fletcher, Chief Information Security Officer at Blackstone, described the reasoning behind the lead investment: “​​The way organizations secure internet-facing applications is fundamentally changing. AI-driven traffic and increasingly fragmented edge environments require a new approach to security management. Huskeys has built a platform designed to operate at enterprise scale while advancing a new category for the modern network edge. We believe the company is well positioned to redefine how organizations manage edge security, and we’re excited to support Huskeys as they continue to develop their product in this important category.”

Customers That Monetize Their Traffic

The reference list maps closely to the revenue argument. Huskeys’ customer base includes TikTok, LEGOLAND, Ro, Blackstone, and Hugging Face.

TikTok monetizes attention at extraordinary volume. LEGOLAND sells tickets and experiences with a direct link between site availability and gate revenue. Ro operates a regulated healthcare business where trust in the transaction is part of the product. Hugging Face serves developers and automated systems that pull models and datasets continuously. Blackstone brings the requirements of a major financial institution and is now also the lead investor.

Across those accounts, Huskeys analyzes more than a trillion web requests and thousands of network configurations every day.

The Category Being Funded

Gafni tied the founding decision to the tools security teams were handed: “We founded Huskeys because security teams are being asked to protect increasingly complex edge environments with legacy tools that were never designed to work together or to address the challenges of the new AI era.” He continued: “As AI transforms both legitimate traffic and cyberattacks, organizations need an intelligence layer that continuously understands what’s happening across the edge and adapts security in real time. Our goal is to make the network finally work for the business, not against it. That’s the category we’re building.”

What Boards Will Want to See

The pitch to a finance function is measurable in principle. Fewer legitimate transactions interrupted. Shorter time from vulnerability disclosure to mitigation. Less engineering time diverted to emergency edge changes. More consistent policy across an estate that grew by accumulation.

Whether those numbers materialize consistently across different architectures is the question the next stage of the company will answer. What is already established is the scale at which the answer will be tested. A trillion requests a day, across customers ranging from a social platform to a theme park operator to an AI model hub, is a broad enough sample to make the results meaningful.

Subscribe now

Subscribe now

Oren Blank’s argument is worth taking seriously on its own terms before getting into the product. The CEO of Walnut, announcing an enterprise-grade agent platform for product-led buyer enablement, framed the release around a claim about where advantage in software has moved:

“Every company has access to the same frontier models now, so the ability to build features isn’t a competitive advantage anymore. What is: your ability to go to market effectively, tell a coherent story, and create a buyer experience that actually helps your buyer make the informed decision.”

If that premise holds, the go-to-market function becomes the place where differentiation is won or lost, and the tools that function uses matter more than they used to. Walnut’s release is a bet on exactly that.

What Walnut Announced

Walnut, the platform used by go-to-market teams at hundreds of B2B software companies to build interactive product demos, deal rooms, and buyer-facing content, has introduced a family of AI agents that covers both sides of the buying experience for the first time. Two agents are new. AI Mode for Playlists and Deal Rooms assembles a complete, personalized buyer-facing experience from a single prompt. Walnut Xpert sits on top of a demo and lets buyers ask questions and find their own way through the product. They join InsightsAI and AI Mode for demos, already live on the platform.

Coherence as a Product Requirement

Blank’s quote emphasizes telling a coherent story. The architecture of the platform is designed around that requirement. All four agents share the same foundation. Rather than generating content from a prompt and a guess at what the product looks like, Walnut’s agents work from the company’s own capture technology, which clones the actual product, its real screens, real flows, and real data, and builds every asset on top of that captured context.

The result is that when AI Mode assembles a demo or Walnut Xpert answers a buyer’s question, the output is anchored to what the product actually does. Walnut describes this as the architectural bet behind the announcement: a buyer’s journey holds together from first click to signed deal because every agent along the way draws on the same grounded context rather than starting from a blank page. A deal room assembled by AI Mode reinforces the story the demo already told rather than starting over.

Personalization at Full Scale

The second half of the strategy is reach. In Walnut’s telling, personalization stops being something a team builds for its five biggest deals and becomes something it builds for all of them.

Two data points support the claim. Deal room and playlist curation, which used to take an account executive or customer success manager roughly three hours per deal, now takes a sentence. Demo creation, over the past year of Walnut’s demo creation and personalization agents, has moved from days to hours to minutes, with greater fidelity and interactivity rather than less. A GTM team that once had to choose which accounts merited a tailored experience can now extend that experience across the pipeline.

The Buyer as a Source of Signal

Blank’s third clause, a buyer experience that helps the buyer make an informed decision, is where Walnut Xpert enters. Instead of a fixed, click-through path, a buyer can ask questions, go deeper on what matters to them, and find their own way through the product, guided by an agent grounded in the seller’s actual captured product.

Walnut reports that sessions with Walnut Xpert alongside a demo more than double the time a buyer spends. The questions asked generate intent signal a static demo never could. Walnut compares the effect to an AI SDR, qualifying interest and surfacing what buyers care about at a scale no team could staff for. Awareness activity that once yielded a click or a page view becomes a record of what buyers actually asked, in their own words.

AI Mode for Playlists and Deal Rooms in Practice

A seller describes what a specific buyer needs. The agent assembles a complete, personalized playlist or interactive deal room, drawing on the company’s captured product content, the buyer’s CRM record, and the seller’s stated intent. Because it works from the same grounded foundation as every other Walnut agent, the output reinforces rather than restarts the story.

The Model Question

Blank’s opening premise, that every company has access to the same frontier models, also explains Walnut’s decision to support bring-your-own-key (BYOK). Customers will be able to use their own model key, gaining direct control over security, privacy, and spend. Walnut’s position is that its verticalized buyer-enablement harness, the captured context, the workflows, and the governance, is the durable value, and it works with any underlying model. As frontier models keep improving, the quality and value customers get from Walnut’s agents improves with them, regardless of which model is doing the work underneath.

Governance as the Precondition

Before any agent with autonomy shipped, Walnut released the AI Center, a governance layer that gives administrators visibility, permissions, and guardrails over how AI is used across their workspace. The sequencing was deliberate. Agents that act on an organization’s behalf are only adoptable once someone inside that organization can see and control what they are doing, across every agent, in one place. BYOK support extends that layer.

Judgment Stays Human

In every case, the person on the GTM team remains the one who decided what story should be told and to whom. The agents execute from one shared, grounded context. The judgment stays with the humans who understand the account. That division is consistent with Blank’s framing: the strategic act of going to market well remains a human act, and the agents are there to carry it out at scale.

Availability

The AI Center, AI Mode for demos, AI Mode for Playlists and Deal Rooms, and Walnut Xpert are available to Walnut customers today. BYOK support, along with a companion MCP that lets customers run it from the AI provider of their choice, will roll out in the coming weeks.

Walnut describes itself as a vertical AI platform for product-led buyer enablement, built on a single agentic architecture that lets GTM teams direct one cohesive buying story instead of stitching together disconnected AI tools. On Blank’s reading of the market, that cohesive story is the thing worth competing on.

Subscribe now