Trace has raised $3 million to solve the problem of enterprise adoption of AI agents.

For all their potential, AI agents have been slow to make an impact in the enterprise, and one emerging startup speculates that the reason they haven’t is because of a lack of context.

Trace, launched as part of Y Combinator’s Summer 2025 cohort, is a workflow orchestration startup that aims to fill this gap. The company maps complex enterprise environments and processes to ensure agents have the context they need to scale quickly.

“OpenAI and Anthropic are building a great pool of interns that we can leverage within our company,” said Trace CEO Tim Cherkasov, referring to the AI ​​Lab’s tools. “We’re building managers who know where to put them.”

The London-based company said on Thursday it had raised $3 million in seed funding from Y Combinator, Zeno Ventures, Transpose Platform Management, Goodwater Capital, Formosa Capital and WeFunder. Angel investors Benjamin Bryant and Kevin Moore also invested.

Trace’s system begins by building a knowledge graph from a company’s existing tools – systems like email, Slack, and Airtable – that shape the company’s daily work life. Given that context, users can message the system with high-level tasks such as “We need to design a new microsite” or “Let’s develop a sales plan for 2027,” and Trace provides a step-by-step workflow, delegating some tasks to AI agents and assigning others to human workers. When the system calls an AI agent, it is prompted for specific data needed to complete a subtask.

The idea is to automate the delicate task of onboarding AI agents, one of the biggest obstacles to real-world deployment within companies.

With so many companies focused on agent AI, Trace will have a lot of competition. Earlier this week, Anthropic launched its own take on enterprise agents, focusing on pre-built plugins for specific departmental functions. And many of the workplace productivity services that Trace will leverage, like Atlassian’s Jira, are rolling out their own agents to potentially compete with startup systems.

However, Trace’s founders believe that the knowledge graph approach will be key to success as it allows context engineering to be built deep into the agent deployment structure.

“2024 and 2025 were still the era of prompt engineering. Now we have moved from prompt engineering to context engineering,” says CTO Arthur Romanov. “Whoever can provide the best context at the right time will be the infrastructure on which AI-first companies will be built. And we hope to be that infrastructure.”