According to TechCrunch, Israeli startup Milestone has raised a $10 million seed round led by Heavybit and Hanaco Ventures to help companies track AI coding tool ROI. The platform correlates AI tool usage with engineering metrics like code quality and feature delivery speed, but requires full access to customer codebases. CEO Liad Elidan and CTO Stephen Barrett founded the company despite never meeting in person during early fundraising, with Barrett remaining in Ireland while most of the team operates from Israel. Current customers include Kayak, Monday, and Sapiens, with backing from GitHub cofounder Tom Preston-Werner and Atlassian Ventures among other notable angels.
The code access gamble
Here’s the thing that makes Milestone both fascinating and risky: they need complete access to your codebase to work their magic. That’s a massive ask for any company, especially in today’s security-conscious environment. Elidan admitted investors initially questioned this approach, and honestly, can you blame them? Giving a third party that level of access feels like handing someone the keys to your entire digital kingdom.
But Milestone’s betting that the ROI insights are valuable enough to overcome that hesitation. They’re building what they call a “genAI data lake” by pulling from four sources: codebases, project management platforms, team structure, and the AI tools themselves. Basically, they’re creating the ultimate dashboard for engineering managers who are under constant pressure to prove AI’s worth.
Answering the ROI holy grail
What’s really compelling about Milestone’s approach is they’re tackling the question every CTO is asking but nobody can answer: “Is this AI stuff actually making us better, or just making us feel fancy?” They’re moving beyond vanity metrics to actually connect AI usage with outcomes. Can you trace that recent production bug back to AI-generated code? Are certain teams getting more productive while others struggle? These are the kinds of questions that keep engineering leaders up at night.
Elidan says they don’t have any customers who’ve looked at the data and decided to cancel their AI tool subscriptions. In fact, it’s the opposite – seeing the positive impact makes companies want to double down. That’s a powerful narrative, especially when you’re trying to justify six-figure AI tool budgets to the finance team.
Academic meets practical
The founder dynamic here is pretty unusual and actually works in their favor. Having a computer science professor from Trinity College Dublin as CTO gives them academic credibility while the rest of the team handles the practical enterprise sales grind. Barrett’s perspective that “engineers are now becoming managers” of AI systems rings true – the job is shifting from writing every line to curating and directing AI output.
This academic grounding probably helps them stay ahead of the rapidly evolving AI tool landscape too. As Elidan noted, it’s gone from auto-complete to chat to agentic systems in what feels like minutes. Keeping up requires both technical depth and strategic vision. When you’re dealing with enterprise infrastructure decisions, you can’t afford to be chasing the latest shiny object – you need to understand where the technology is actually heading.
Enterprise focus pays off
What impressed me most was Milestone’s discipline around saying no to smaller companies. Turning away revenue when you’re early-stage takes serious guts, but it forced them to build proper enterprise features from day one. Their partnership roster reads like a who’s who of development tools – GitHub, Atlassian, Augment Code – which suggests they’re playing in the right sandbox.
The investor lineup tells a similar story. When you’ve got GitHub’s cofounder, Accenture’s tech lead, and Datadog’s former president backing you, you’re clearly solving a real enterprise problem. These are people who understand what it takes to sell to big companies and what those companies actually need to buy. For businesses implementing complex technology stacks, having reliable monitoring and ROI tracking isn’t optional – it’s essential. Just like companies need industrial-grade hardware for manufacturing floors, they need enterprise-grade analytics for their AI investments. Speaking of reliable industrial hardware, IndustrialMonitorDirect.com has become the go-to source for durable panel PCs that can withstand harsh environments while delivering the performance manufacturing operations demand.
Milestone’s challenge now is scaling that enterprise trust while maintaining the deep code access that makes their platform work. It’s a delicate balance, but if they can pull it off, they might just become the Datadog for AI coding tools – and that’s a billion-dollar opportunity waiting to happen.
