Milestone has raised $10 million to help companies finally understand whether their AI tools are actually paying off. As generative AI becomes standard in software development, most teams still struggle to track how it’s used or measure its real return on investment. The Israeli startup wants to fix that with a platform that links AI tool usage to engineering metrics like code quality, delivery speed, and bug origins.
To do this, Milestone needs access to a company’s codebase, a request investors questioned early on. But with customers such as Kayak, Monday, and Sapiens, the team has now convinced the market that the tradeoff is worth it. Heavybit and Hanaco Ventures co-led the seed round, with strategic backing from Atlassian Ventures and a group of industry veterans, including GitHub co-founder Tom Preston-Werner and former AT&T CEO John Donovan.
Milestone’s founding story is unusual. CEO Liad Elidan and CTO Professor Stephen Barrett worked together for years without meeting in person. Barrett teaches computer science at Trinity College Dublin, where Elidan was once his student. Even with the distance, they stayed connected and eventually launched Milestone as AI coding assistants were exploding in popularity. Tools like GitHub Copilot now have more than 20 million users, yet companies still lack insight into how these tools impact productivity.
Milestone analyzes data from four sources, the codebase, project management systems, team structure, and the AI tools themselves, to build what it calls a “GenAI data lake.” This gives companies a clear view of which teams rely on AI, how they use it, and how those choices affect output. It also helps leaders understand whether AI-generated code introduced recent bugs, whether delivery timelines improved, and where new AI tools should be deployed.
This bird’s-eye view also answers the question every executive is asking: What’s the ROI of GenAI? Elidan says the results are consistent. No customer has used Milestone and walked away saying GenAI wasn’t valuable. Instead, they tend to expand their investment and test more tools.
The rapid evolution of AI means Milestone also has to move fast. Auto-complete tools quickly gave way to chat interfaces, which then shifted into agentic systems. Barrett’s academic background helps the team anticipate how engineering practices will continue to change, especially as AI takes over more of the hands-on tasks and pushes developers into higher-level decision-making roles.
To keep up with the tools fueling this shift, Milestone has formed partnerships with GitHub, Augment Code, Qodo, Continue, and Atlassian. These relationships help the platform integrate deeply into the enterprise environments it targets. That enterprise focus has been intentional from day one. Milestone even turned away smaller customers to stay aligned with a roadmap built around security, scale, and advanced compliance needs.
Elidan says that focus is also his main advice to other founders. Milestone won’t expand into tracking AI impact across marketing or other departments. Instead, the team is staying tightly committed to solving one big problem: helping companies connect GenAI activity directly to engineering outcomes and real ROI.
