SAN FRANCISCO, CA — Momentic has launched a major platform update that rethinks software verification for the AI era. AI coding tools have made it possible for teams of all sizes to ship more code than ever before, but how that code gets validated has not kept up. The result is bugs reaching production faster than teams can catch them, and engineering teams stuck firefighting hotfixes instead of shipping.
Every Momentic agent — whether building new tests, classifying failures, or healing broken ones — runs on a shared knowledge base of how your product actually works. Teams can define terminology, critical flows, and common failure patterns; Momentic maps every user journey and keeps that understanding current as the product evolves. It ingests Jira tickets, GitHub, support docs, Slack, and connects with other internal tools, so the system always knows what’s changing, what’s breaking, and whether a failure is a real bug or from an intended change. The result is a system that gets smarter the more you use it.
Key capabilities in this release include:
Test Coverage That Grows With Your Product
Most teams share the same problem: coverage gaps that widen every sprint because writing new tests doesn’t scale with the speed of shipping. Momentic’s Explore Agent closes that gap automatically. It monitors pull requests, reads diffs and codebase context, identifies which flows lack coverage, and proposes new tests or updates existing ones, without any manual authoring.
Failure Classification and Healing: From Noise to Signal
Flaky tests and noisy CI pipelines are one of the biggest drains on engineering time, and one of the main reasons teams stop trusting their test suites. Momentic’s Failure Classification Agent automatically categorizes every failure: real bug, intended application change, test setup issue, or transient error. When a failure isn’t a real bug, Momentic proposes a fix and opens a PR. When it is, the team gets a high-signal alert with full context on exactly what broke and why.
Legacy test frameworks produce artifacts that only the engineer who wrote them can understand. Momentic’s new test format is intent-based and readable by both humans and AI agents. Engineers describe what they want to test in plain English; Momentic executes it live against a browser or emulator.. AI agents can parse, build, and modify tests more effectively, making the entire development loop faster and more autonomous.
Momentic’s customers are already seeing the benefits of its release with measurable impact. In a few weeks of beta access alone, Momentic’s agents analyzed over 70,000 test failures, opened 400 pull requests, and wrote 600 tests with a 73% merge rate. In total, customers have executed more than 2 billion steps on the platform and verified over 80,000 PRs: proof that agentic verification doesn’t just keep up with the AI coding tools, it actually gives engineers the confidence to ship more with them, faster.
For more information about Momentic’s availability and pricing, visit momentic.ai/pricing.
