Neo has $100 million and a sharp claim: enterprise security still can't tell when software is acting like an employee.
Neo emerged from stealth on July 20, 2026 with $100 million in seed and Series A funding from Andreessen Horowitz, Bessemer Venture Partners, Craft Ventures and Merlin Ventures. SecurityWeek reported that the Boston-based company is building a control layer for AI agents and AI-enabled software running inside enterprise environments. That is the bet. If software can now reason and invoke tools, moving through workflows with valid user permissions, the old security stack has a visibility problem before it has a malware problem.
The founding team gives the pitch its weight. Nick Warner, Neo's CEO, previously served as SentinelOne's president and chief operating officer and helped take the company public in 2021. Shlomi Salem, Neo's chief product officer, previously led detection engineering at SentinelOne. Eran Shirazi, the company's chief technology officer, co-founded EasySend, according to Neo's launch announcement. These aren't outsiders looking for a fashionable AI angle. They helped build or scale the kind of enterprise software buyers already know.
The blind spot is identity
Here's the gap Neo is chasing. AI agents can operate inside browsers and developer tools, and across SaaS platforms, using real employee credentials. To an identity system, that agent can look like the employee. To an endpoint tool, the activity can look like normal application behavior. But the thing clicking buttons, pulling data or chaining tools may not be a person at all.
Neo's platform, according to the company announcement distributed through GlobeNewswire, keeps an inventory of AI agents, AI-enabled applications, browser tools, extensions and other software with agentic capabilities. It assesses what those tools can access, whether they're configured safely, and which user, agent or application is responsible for an action. That's the inventory part. It also lets security teams set policy for tool calls, API access, data movement and agentic workflows before risky activity runs loose. That is a different job from watching a laptop for known malware.
You can see why investors moved quickly. Gartner forecast that the share of enterprise applications with task-specific AI agents will rise from less than 5% in 2025 to 40% by the end of 2026. That isn't a distant platform shift. It is software vendors adding autonomy into products companies already approved, while employees bring in their own agents from the bottom up.
Frankly, that makes the risk awkward for security teams. Nobody wants to block every AI tool and become the department of no. But nobody serious can pretend an agent with access to Salesforce, GitHub, Slack or a browser session is just another harmless productivity feature. Once software acts through a trusted user, the audit trail has to say more than which account was logged in.
A security category getting crowded fast
Neo isn't walking into an empty market. The Wall Street Journal described AI security startups such as Armadin, 7AI and Tenex as part of the same investor rush around agentic software risk. Neo's own launch materials also point to a team with experience from SentinelOne and Wiz, plus Palo Alto Networks. That matters because security categories harden quickly once large enterprises decide a new budget line is real.
Wiz is the obvious recent example. Google said in March 2026 that it completed its $32 billion acquisition of the cloud security company, after announcing the deal the previous year. Cloud security did not stay a niche for long once workloads moved out of corporate data centers. Agentic software security is trying to tell the same story, only faster.
The harder question is whether Neo becomes a platform or just one more tool security buyers have to stitch together. Neo says it will use the capital to grow engineering and go-to-market teams. That is necessary, but it isn't enough. Enterprises will ask whether Neo can sit across identity, browsers, SaaS applications and developer workflows without slowing the work it is supposed to secure.
That is where Warner's SentinelOne past cuts both ways. It gives Neo credibility with security buyers who remember the endpoint wars. It also raises the standard. If you tell enterprises their existing controls were built for a more predictable software world, you have to show them the new control point clearly, not just name the category before everyone else does.
Neo has timing and money - and founders who understand the buyer. Now it has to prove that agentic software control is a product companies need this year, not a label investors liked in July 2026.
Also read: Gritt Robotics Raises $34 Million to Automate Solar Farm Construction • How to Split Equity Between Co-Founders Without Blowing Up the Company • Natural raised $30 million to give AI agents their own bank accounts
