Baz Raises $9M to Push AI Coding From Review to Planning Stage
The Israeli agentic coding platform also introduced a new product, which intervenes before code is written to detect and prevent bugs
Illustration: Jakub Zerdzicki via Pexels.com
Baz, an Israeli agentic coding platform that helps engineering teams detect and prevent software bugs before code is written, has raised an additional $9 million in an extended seed round. The round brings the company’s total funding to $17 million and was co-led by Battery Ventures and boldstart ventures, with participation from AFG Partners and Disruptive VC.
This announcement comes alongside the launch of Baz Planner, a new product the company says shifts engineering workflows further upstream – from reviewing code after it’s written to shaping how it is planned in the first place.
Instead of waiting for code changes to be flagged post-facto, Baz Planner sits between developers and the codebase, routing proposed changes through automated analysis loops. It evaluates ideas against architecture, security, and reliability constraints, flags risky paths, and can block or redirect changes before they are implemented.
Baz says early users have seen up to a 65% reduction in downstream rework, based on fewer rollback and hotfix pull requests after code merges. The company positions itself as infrastructure for the AI coding era, where large portions of code are generated by machines rather than written manually.
“Our customers are building the backbone of the new AI stack, and they pushed us to go beyond review and intervene earlier in the process — at the planning stage, where bugs and vulnerabilities are cheapest to eliminate,” said CEO Guy Eisenkot.
Baz’ core product, AI Code Review, is already used by over 100 companies across AI, infrastructure, and cybersecurity. The platform uses specialized agents that act across different layers of the engineering stack, including product requirements, security, system reliability, and architecture.