Automated Testing

AI for SOX testing. Midship is rebuilding internal audit from the ground up. We're starting with SOX testing by building AI agents that perform audit work end to end, giving auditors leverage to focus on higher-value risk and advisory work.

Replace the humans fixing bad data. Mica's AI agents replace the data ops teams fixing bad data. When bad or missing data breaks the pipeline, and orchestration, retries, and monitoring fail, painful manual review work kicks in, pulling humans in to investigate and patch data issues across systems. Mica does what those humans do: gathering the right information from internal docs and external systems, reasoning across context, and resolving errors autonomously to get the pipeline moving again. The result: dramatically reduce time, cost, and operational drag as your data pipelines scale without scaling ops headcount. Mica turns judgment-heavy data fixes from a manual bottleneck into an automated background process with full auditability.

QA agents that write tests 10x faster with zero maintenance. Decipher AI generates your E2E tests, maintains them as you ship, and catches issues before customers do. Plus, get alerted when users hit live bugs. Used by the fastest shipping companies including Bilt, Arize, and Vial.

AI code review agent with complete context of your codebase. Greptile reviews pull requests with full context of the codebase. Top YC companies like PostHog, Raycast, Vouch, Podium, and YC's own internal software team use Greptile to review millions of changes every week. When companies start using Greptile, they catch more bugs and merge PRs 4x faster, on average Greptile can be used via the cloud, or be hosted 100% on your cloud in an air-gapped VPC.

Tools for agent verification. Remember the first time you hit and your code actually worked? Something that didn't exist before, now alive on your screen. Today, coding agents can do that, orders of magnitude faster. But their verification is shallow at best. The faster agents generate code, the wider the gap grows. Nothing verifies software at the speed agents ship code. We believe verification unlocks long-horizon coding agents: agents that run for days and prove their work as they go.

AI agent that generates unit and integration tests. Tusk (YC W24) is an AI agent that helps developers generate unit and integration tests with codebase context. Tusk sits in CI and provides automatic coverage whenever software engineers raise a PR. Engineering teams use Tusk to prevent critical regressions while increasing test coverage.