/Tech5h ago

Momentic launches AI-powered testing agents that ingest PRDs and PR diffs to automate software QA

Story Overview

Momentic is rolling out AI agents that pull product requirements and code changes straight from tools like Notion and PR diffs, then write, update, and triage tests without rigid scripts. The update also adds a shared knowledge base for team standards and a plain-English test format meant to keep humans and agents on the same page while AI coding speeds up.

113959347559.4K

#1257

Original post

Santiago@svpino#1889inTech

I sent this to the entire team I’m currently working with.

It’s a very good team, but they are still on a 10-year-old process with project managers that can’t read code, specialized developers, and a full QA team.

I don’t think you can afford writing software using a 10-yo playbook.

wei-wei@wuweiweiwu

AI writes all your code. Does it still work?

That's the gap: coding agents are scaling output, but the same with bugs and incidents. The biggest bottleneck is QA: how can you make sure the code your agents just wrote doesn’t break everything?

The best QA are the ones who know everything about your product. Today, we give that knowledge to Momentic’s most capable agent.

Every Linear ticket, every PRD in Notion, every PR - Momentic knows all.

In just the past few weeks, our agents analyzed 70k+ test failures and created 600 tests. Achieving a 73% PR merge rate, and it’s growing fast.

We’re grateful for the trust of our customers, including Notion, Xero, Webflow, Retool, Runway, Bilt, and more.

9:07 AM · Jun 23, 2026 · 551.6K Views

Developer Impact

Agents now handle triage and coverage on their own

Instead of engineers chasing every failure, the new Failure Classification Agent sorts real bugs from intentional tweaks and can open its own PRs to fix tests. An Explore Agent scans diffs to grow test coverage around new features and edge cases.

Open Question

Self-serve access is live but long-term results stay unclear

Teams can jump in today via a simple CLI command and try the refreshed platform for free, yet the announcement leaves open how the agents perform once code velocity keeps climbing across larger codebases.

Sentiment

Positive users are excited that Momentic's AI agents could transform QA with product context from tickets and PRDs, while negative users complain the agents generate unwanted tests, miss tacit knowledge, and increase debugging time.

Pos

25.0%

Neg

75.0%

3 comments with sentiment.

Cluster Engagement

Digg Deeper

No Digg Deeper questions have been answered for this story yet.

MOMENTICVia

#1257

Posts from X

Most Activity

VIEWS4.6KBOOKMARKS12LIKES12RETWEETS4REPLIES4

Rohan Paul@rohanpaul_ai

AI coding has a QA problem.

And Momentic just released a testing platform around that bottleneck: faster code means little if QA cannot scale with it.

So Momentic's testing agents read the same product memory humans use, including Linear tickets, Notion PRDs, and PR diffs, so tests are based on intended behavior rather than blind script matching.

They are claiming in the past few weeks, its agents analyzed 70K+ test failures, created 600 tests, and reached a 73% PR merge rate.

Meer | AI Tools & News@Meer_AIIT

AI is writing more code than ever, but the bugs are scaling right along with it. momentic just put out how they're closing that gap. the core problem they point to: → coding agents ship faster

→ but more output means more bugs and incidents

→ qa becomes the real bottleneck their take is that the best qa knows everything about your product. so they gave that knowledge to their agent. what the agent reads: → every linear ticket

→ every prd in notion

→ every pr the numbers from the past few weeks: → analyzed 70k+ test failures

→ created 600 tests

→ hit a 73% pr merge rate, and growing customers they named include notion, xero, webflow, retool, runway, and bilt.

5h4.6K1212

ben hylak@benhylak

looking good @wuweiweiwu

wei-wei@wuweiweiwu

AI writes all your code. Does it still work?

That's the gap: coding agents are scaling output, but the same with bugs and incidents. The biggest bottleneck is QA: how can you make sure the code your agents just wrote doesn’t break everything?

The best QA are the ones who know everything about your product. Today, we give that knowledge to Momentic’s most capable agent.

Every Linear ticket, every PRD in Notion, every PR - Momentic knows all.

In just the past few weeks, our agents analyzed 70k+ test failures and created 600 tests. Achieving a 73% PR merge rate, and it’s growing fast.

We’re grateful for the trust of our customers, including Notion, Xero, Webflow, Retool, Runway, Bilt, and more.

4h2.1K120

Rohan Paul@rohanpaul_ai

Read their technical blog on this

https://momentic.ai/blog/a-new-era-of-software-quality

Rohan Paul@rohanpaul_ai

AI coding has a QA problem.

And Momentic just released a testing platform around that bottleneck: faster code means little if QA cannot scale with it.

So Momentic's testing agents read the same product memory humans use, including Linear tickets, Notion PRDs, and PR diffs, so tests are based on intended behavior rather than blind script matching.

They are claiming in the past few weeks, its agents analyzed 70K+ test failures, created 600 tests, and reached a 73% PR merge rate.

5h1.2K21

wei-wei@wuweiweiwu

@svpino 🙏🏼🙏🏼🙏🏼 we’re going to blow their minds

5h301

wei-wei@wuweiweiwu

@benhylak The momentic peptides be hitting 🤣

4h4

Thomas Tao@Thomas_Tao_1

@rohanpaul_ai Yeah. The ugly part is triage. More generated code means more noisy failures, fast.

4h4

James Clawn@JamesClawn

@rohanpaul_ai Notion PRDs are useful context until they drift. The QA agent probably needs a trace showing which requirement version shaped each failed test.

5h3

Bojan@bojan_ai

@rohanpaul_ai This is the real bottleneck nobody posts about. AI writes the feature in 10 minutes, then I spend 2 hours confirming it didn't quietly break three other things. The bug count didn't drop, it just moved downstream to review.

3h2

FemMiraMancer@miratechgeek

@svpino Agents give unwanted tests! I don't want to debug. They read Linear. Miss tacit knowledge. 73% merge rate. I fix the rest. QA isn't dead. The QA is me, and I'm annoyed.

5h2

Quixoa@Quixoa_

@svpino This is where a lot of AI rollout effort gets wasted. Teams add copilots and agents on top of a process that was already built around handoffs, queues, and translation layers. The workflow usually has to get narrower before AI makes it faster.

4h1