Autoheal raises $7.9M to evaluate and fix AI agents with… AI agents

Autoheal AI Inc., a startup focused on developing AI-native platform engineering solutions, has raised $7.9 million in seed funding. The funding, led by Innovation Endeavors and supported by other venture capitalists, will help the company create a unified platform for building, governing, and improving AI agents that operate within the software development lifecycle.

Enterprises are increasingly using AI tools to generate code, but this has led to a rise in production incidents and software vulnerabilities. Autoheal's platform provides a secure environment for managing AI agents, ensuring they have access to the same engineering context, cost controls, and security measures. The platform includes two specialized agents - an "Evaluator agent" that rates downstream worker AI agents based on metrics like CI failures and incident reports, and a "Healer agent" that fixes low-scoring agents by opening pull requests to improve model selection, prompts, tools, and skills.

Autoheal's platform has already seen adoption from several enterprise customers, including Normura Holdings Inc., AvidXchange Inc., and Empiric Earth Inc., who have reported reduced incident resolution times and saved engineering hours. The startup aims to develop reinforcement learning techniques to train customers' AI agents on private engineering data, enabling them to create enterprise-specific small language models that operate within their secure private cloud environments.

Ultimately, Autoheal envisions expanding its architecture beyond software engineering to data and security engineering domains as well.