Trace Secures $3M to Boost AI Agent Adoption in Enterprises
▼ Summary
– Trace, a new Y Combinator startup, believes AI agents have underperformed in enterprises due to a lack of context and aims to solve this with workflow orchestration.
– The company builds a knowledge graph from a firm’s existing tools (like Slack and Airtable) to provide AI agents with the necessary context to scale.
– Users can give Trace a high-level task, and it will generate a step-by-step workflow, delegating tasks between AI agents and human workers with specific data.
– Trace recently raised $3 million in seed funding but faces competition from other enterprise AI agent solutions, including new offerings from Anthropic and productivity platforms.
– The founders argue the industry is shifting from “prompt engineering” to “context engineering,” and they aim to be the foundational infrastructure for AI-first companies.
While AI agents hold immense promise for transforming business operations, their widespread adoption within enterprises has faced significant hurdles. A primary challenge lies in the lack of contextual understanding these agents possess about a company’s unique environment and processes. Trace, a new workflow orchestration startup, aims to solve this problem by providing the crucial context AI needs to function effectively at scale. Emerging from Y Combinator’s 2025 summer cohort, the company recently secured $3 million in seed funding to advance its platform.
The funding round was led by Y Combinator and included participation from Zeno Ventures, Transpose Platform Management, Goodwater Capital, Formosa Capital, and WeFunder. Angel investors Benjamin Bryant and Kevin Moore also contributed. The London-based startup positions itself as the essential layer between powerful AI models and practical enterprise application. CEO Tim Cherkasov describes the dynamic by comparing leading AI tools to brilliant interns. “We’re building the manager that knows where to put them,” he explains, emphasizing Trace’s role in directing and deploying AI capabilities intelligently.
Trace’s system operates by first constructing a detailed knowledge graph. This map is built by integrating with a company’s existing suite of tools, such as email clients, Slack, Airtable, and project management software, that define daily workflows. Once this contextual foundation is established, users can assign high-level objectives. For example, a manager might instruct the system to “design a new microsite” or “develop our 2027 sales plan.” Trace then generates a complete, step-by-step workflow. It intelligently delegates tasks, assigning some to AI agents and others to human team members. Critically, when an AI agent is activated, the system provides it with the specific data and context required to complete its assigned sub-task accurately.
This approach directly tackles one of the major barriers to enterprise AI adoption: the complex and delicate process of onboarding and integrating these agents into existing systems. By automating this integration and providing deep contextual awareness, Trace seeks to unlock the practical utility of agentic AI. However, the competitive landscape is intensifying rapidly. Many companies are now focused on developing agentic solutions. Just this week, Anthropic launched its own suite of enterprise agents designed for specific departmental functions through pre-built plugins. Furthermore, many of the productivity platforms Trace connects to, like Atlassian’s Jira, are beginning to launch their own native AI agents, creating potential competition for the startup’s orchestration layer.
Despite this competition, Trace’s leadership is confident that their foundational technology provides a distinct advantage. The company believes the next phase of AI deployment hinges on context, not just prompts. “2024 and 2025 was still about prompt engineering. Now we’ve moved from prompt engineering to context engineering,” states CTO Artur Romanov. He argues that the winning infrastructure for AI-driven companies will be whoever can deliver the best context at the precise moment it’s needed. “We hope to be that infrastructure,” Romanov concludes, framing their knowledge-graph methodology as the key to enabling truly effective and scalable AI agents in the corporate world.
(Source: TechCrunch)
