NeoCognition, an AI agent research lab founded by Ohio State University professor Yu Su, emerged from stealth on April 21 with $40 million in seed funding. The oversubscribed round was co-led by Cambium Capital and Walden Catalyst Ventures, with participation from Vista Equity Partners. Angel investors include Intel CEO Lip-Bu Tan and Databricks co-founder Ion Stoica.
The thesis is blunt: current AI agents are unreliable. Su told TechCrunch that agents from Claude Code, OpenClaw, and Perplexity’s computer tools complete tasks as intended only about 50% of the time. “Every time you ask them to do a task, you take a leap of faith,” Su said.
The Specialization Bet
NeoCognition’s approach mirrors how humans develop expertise. Rather than building general-purpose agents that attempt every task with the same broad capability set, the company is developing agents that autonomously learn the structure, workflows, and constraints of specific domains.
“The true power of human intelligence is the ability to continuously learn and specialize,” Su told TechCrunch. “Our approach mirrors how humans gain expertise on the job through building a structured model of their micro-world.”
The company’s founding research team contributed seminal work in the agent field, including Mind2Web, MMMU, and SeeAct, according to the company’s press release. Research from Su’s Ohio State lab is used in frontier models from OpenAI, Anthropic, and Google.
The Investor Logic
The backing from Vista Equity Partners, one of the largest private equity firms in the software space, signals a clear enterprise distribution strategy. Vista’s portfolio of SaaS companies gives NeoCognition direct access to enterprises looking to embed agent capabilities into existing products.
“As general-purpose agents become table stakes, the key challenge in AI is achieving expert-level intelligence,” Stoica said in the announcement. “NeoCognition’s new approach to building agents that learn to become experts has the potential to reach the level of reliability, efficiency, and cost-effectiveness required for high-stakes applications.”
The Reliability Premium
NeoCognition currently has about 15 employees, the majority holding PhDs. The company is headquartered in Palo Alto.
The $40M seed values a bet that the market for agent reliability tooling is about to expand sharply. InsightFinder AI closed a $15M Series B earlier this month targeting agent failure detection. Nava raised $8.3M for agent trust infrastructure in financial transactions. NeoCognition is going a layer deeper: rather than catching failures after they happen, the pitch is that agents should learn enough about their operating environment to stop failing in the first place.
The difference between a 50% success rate and a 95% success rate is not incremental improvement. It is the threshold between a tool that requires constant supervision and one that can be trusted to operate autonomously. That is where the $40M is pointed.