As companies start using artificial intelligence in real products, a new challenge is appearing: ensuring these systems remain reliable after launch.
AI tools are now managing customer interactions, internal processes, and key features. However, when issues arise, it can be hard to figure out the cause. San Francisco–based Braintrust says it wants to solve that challenge.
The company has raised $80 million in a Series B funding round led by ICONIQ. Existing investors Andreessen Horowitz, Greylock, Elad Gil, and Basecase Capital also participated.
With the new funding, the US company plans to expand its engineering and go-to-market teams, open additional offices, and develop new products.
Observability layer for AI
Braintrust positions itself as an observability platform for production AI systems. In simple terms, it helps companies monitor, evaluate, and understand how their AI models and agents perform after deployment.
Founder Ankur Goyal said the idea for Braintrust came from his own experience. While working at Impira and Figma, he had to build internal tools to evaluate AI systems.
The process was complex and time-consuming. After facing similar challenges more than once, he concluded that other teams were likely dealing with the same issue.
Since Braintrust was founded, AI adoption has accelerated. What were once prototypes or demos are now embedded into engineering workflows and consumer products. AI agents can run multi-step tasks, call tools, and generate intermediate reasoning steps. These interactions often produce large amounts of data, sometimes hundreds of megabytes per session.
Traditional monitoring tools were not designed for this level of complexity. According to the company, teams often struggle to trace failures across long-running AI processes or to explain unexpected outputs. Braintrust says it even had to build its own database technology internally to manage the scale and structure of AI-related data.
The company argues that observability should now be treated as core infrastructure for AI-powered products. Because AI systems change frequently with model updates and new data, engineering teams need clearer visibility into performance, reliability, and risk.
Several technology companies are already using the platform, including Notion, Replit, Cloudflare, Ramp, and Dropbox.
“At ICONIQ, we have seen that a defining trait among the generational companies is deep, authentic customer obsession. We believe Ankur and the Braintrust team embody this mindset and have been building their product from the start to serve the evolving needs of their customers. The incredible validation we have seen firsthand from leading AI teams using Braintrust is a testament to the depth of their commitment,” says Matt Jacobson, ICONIQ
