A research lab called NeoCognition emerged from stealth today with a $40 million seed round, and the idea behind it is worth understanding if you are thinking seriously about putting AI agents to work in your business.

The oversubscribed round was co-led by Cambium Capital and Walden Catalyst Ventures, with Vista Equity Partners also participating. Notably, Intel CEO Lip-Bu Tan and Databricks co-founder Ion Stoica both came in as angels. When people running chip companies and data infrastructure firms personally back an AI agent startup, that says something about what they think is coming.

What NeoCognition Is Actually Building

Most AI agents today are generalists. You point them at a task and they do a reasonable job, but they do not truly know your business. They do not understand your approval workflows, your specific terminology, the shortcuts your team uses, or the edge cases that your processes have built up over years. Every interaction is essentially from scratch.

NeoCognition’s bet is that the next leap in agent usefulness is not more compute or a bigger model. It is specialisation. The company was spun out of Ohio State University by Yu Su, a Sloan Research Fellow who runs one of the country’s established AI agent labs. Su spent years studying how humans actually become experts in a domain, and his conclusion was that expertise comes from building what he calls a “world model” of the environment you operate in.

For a professional, that means internalising how decisions get made, what the constraints are, which processes can be shortcut and which cannot, and what good output actually looks like. NeoCognition is building AI agents that do the same thing. Rather than being given a task and searching for an answer, these agents continuously learn the structure, workflows, and constraints of the business environment they are deployed in. Over time, they become genuinely domain-specific.

Co-founder Yu Gu described the company’s goal as building agents that “deliver applied research that turns agents into true business value” while also exploring “fundamental questions, not just scaling.”

Why This Matters for Business Owners

If you have deployed AI tools and found they work reasonably well for generic tasks but fall apart on anything that requires domain knowledge, this research direction explains why. Current agents lack a mental map of your specific environment. They can retrieve information and follow instructions, but they cannot reason from an understanding of how your business actually works.

The NeoCognition approach addresses this directly. An agent that builds a world model of your processes can make better decisions with less supervision, handle unusual situations with more confidence, and get genuinely better over time as it learns more about your environment.

NeoCognition plans to sell primarily to enterprises and established SaaS companies that want to build agent-workers into their products. But the underlying capability is what business owners should pay attention to. Agents that truly specialise in a domain are a different category of tool than what most companies are working with today.

What This Means for Business

The broader signal here is that the enterprise AI agent market is maturing past the proof-of-concept stage. When a $40 million seed round is co-led by firms like Walden Catalyst, and backed by the CEO of Intel and a co-founder of Databricks, investors are making a clear statement: domain-specialist agents are the next wave, and the companies that build the right foundations now will be in the best position to take advantage of them.

For businesses, this creates a practical question. Are you building AI capabilities that can absorb and learn from specialist knowledge, or are you bolting generic tools onto existing processes and hoping for the best?

The gap between those two approaches is where AI value is either created or lost. The businesses pulling ahead in AI adoption are not necessarily the ones with the biggest budgets. They are the ones that understand their own data and processes well enough to put an intelligent agent to work on them.

Getting that foundation right, understanding your data, mapping your workflows, and knowing which processes are ready for agent automation, is exactly the kind of work that separates organisations that see real AI returns from those that are still running pilots a year later.

If you are at that stage of building out your AI strategy, that is a conversation worth having now rather than later.

For a deeper walkthrough of tools like this and how they fit together, the free Working With Claude field guide covers the ecosystem end to end. Get the guide.