I’m Yuval, and for the past several years I’ve been building dev tools (some of you might know Amplication). Over that time, we’ve all seen the rise of AI copilots and agents. They are useful for individuals in the IDE, but limited when it comes to how teams actually ship software. I believe the next step is moving beyond copilots into *autonomous, continuous AI* that runs inside the SDLC itself.
That’s what we’re building with https://overcut.ai/. It connects directly to GitHub, Jira, and other tools, and runs workflows that handle things like:
- Reviewing PRs - Writing tests - Generating specs and docs - Triaging and resolving issues - Opening and updating PRs automatically
The difference is that instead of manually operated assistants or black box agents, you define the workflows (like CI/CD pipelines, but for AI agents). This makes it transparent and under your control.
We provide several ready-made playbooks, and you can also build your own for anything you can imagine on top of Git and tickets. There are built-in integrations to GitHub, GitLab, Bitbucket, Azure DevOps, and Jira.
Would love to hear your thoughts: what workflows would you want AI to take over, and what would hold you back from trusting something like this?
Thanks for reading