How autonomous are the agents, and who checks their work?

“A human intervenes at two points: approving the spec before implementation begins, and merging the PR after review. Everything else runs automatically.”

WarpPublished by the organization itself (documentation), no named author

“PostHog is just a little further than that with agents writing about 70% of our PRs, but humans still skim at least 80% of those.”

Jina YoonAuthor, PostHog newsletter · Practitioner, PostHog

“we've merged more than 2.5 million automated maintenance PRs, the vast majority auto-merged with no human in the loop.”

SpotifyPublished by the organization itself (post), no named author

“Customer agent now resolves 72% of support tickets without human escalation.”

HubSpotPublished by the organization itself (earnings call), no named author

“when guests contact us through our AI aid assistant, over 40% of issues are now resolved without a human agent.”

Brian CheskyCo-founder and Chief Executive Officer · Executive, Airbnb

“About 10% of our codes now is committed. That committed is built by agents, autonomous agents out there. Obviously, we check the code before it gets committed.”

Dara KhosrowshahiChief Executive Officer · Executive, Uber

“This is the most important human checkpoint in the loop. An accurate spec leads to a correct implementation; a wrong spec leads to a correct implementation of the wrong thing.”

WarpPublished by the organization itself (documentation), no named author

“The reviewer agent surfaces issues and inconsistencies; the human makes the final call on quality and correctness.”

WarpPublished by the organization itself (documentation), no named author

“Thanks to these autonomous loops, 60%, 60% of UX issues that are identified by either a customer or a salesperson or a CX person or ourselves are fixed within 24 hours.”

Geoff CharlesChief Product Officer (CPO) · Executive, Ramp

“if this is an autonomous diff coming from Minion, we want to give a human reviewer some confidence that this diff has gone through a lot of self-improvement already”

Uday Kiran MedisettyDistinguished Engineer · Practitioner, Uber

“Connect the factory with your GitHub issues, tag the ones you want it to pick up (choose trivial issues and bugs first), run it, and review the resulting PR and the agents' PR comments.”

Noah BlauensteinerAuthor, Xebia articles · Practitioner, Xebia

“One is you start measuring how many human touches you still do to have the agent do the right thing. That's supposed to go down.”

Patrick DeboisProduct DevRel Lead · Practitioner, Tessl

“With this architecture, agentic reasoning is used for intake and context gathering, but every autonomous risk decision is still made by an auditable model, governed under an approved policy.”

Blake WilliamsProduct Manager · Practitioner, Ramp

“Mode · Automation: PR review. Durable session, woken by an external system, often no human in the loop.”

Burke LibbeyPrincipal Engineer · Practitioner, Shopify

“More than one-third of the PRs we merge are now created by agents that run on their own computers in the cloud.”

Cursor (Anysphere)Published by the organization itself (post), no named author

“Thirty-five percent of the PRs we merge internally at Cursor are now created by agents operating autonomously in cloud VMs.”

Cursor (Anysphere)Published by the organization itself (post), no named author

“we have seen around a 20-percentage-point increase in queries resolved without any human intervention, compared to the equivalent non-Gen AI supported Cora journey.”

NatWest GroupPublished by the organization itself (post), no named author

“AskHR answers 94% of common employee inquiries in minutes or less, without human intervention, any time of the day or night.”

IBMPublished by the organization itself (post), no named author

“We are definitely still bottlenecked on reviews, especially for things that are touching some architecture pieces. And it's actually more subtle than just being bottlenecked on review, … we can carve out time differently. It's … bottlenecked on human ability to even fully conceptualize what we're doing.”

Mike KriegerHead of Anthropic Labs · Leader, Anthropic

“actually talked to Claude about the code and say, all right, these are the questions that I would have. Can you go investigate it? So it is kind of … code review, but still human-driven, and for the really important ones. And for the ones that are cosmetic visual changes, it's much more like we'll fix forward if we need to fix forward.”

Mike KriegerHead of Anthropic Labs · Leader, Anthropic

“an autonomous support agent that Zendesk believes will solve 80% of support issues without human intervention.”

Russell BrandomAI Editor · Outside the organization, Zendesk

“You get a pretty high degree of confidence on merge. Today, almost a fifth of all the pull … are merged without any human review or without any human testing. … that is a number that we care a lot about and want to bring higher and higher, of course, within the guardrails of producing really high quality code.”

Daksh GuptaCo-founder · Executive, Greptile

“What we did was we built a system that actually has agents constantly running and creating BAML programs. … We then look at the entire … transcripts, see what tools they use, see what happened, and obviously we as humans can inspect them. But more importantly, we can have agents go inspect them. And agents find what was good, what was bad. … And we can have humans collaborate with these issues to figure out which ones are real, which ones are hallucinations, which ones don't have taste as much as I hate to use that word. And then we can have agents go ahead and create fixes to these problems and go to address them.”

Vaibhav GuptaCo-founder and Chief Executive · Executive, Boundary (BoundaryML)

“But the real problem with all of this is I built this and I hit a little bit of AI psychosis. And I started shipping 10 design docs a day. And soon, the team was fighting my slop. … This last rule was … ship a design doc, you require people to actually go read it. And with this last standard, we suddenly had design docs … incredibly high quality.”

Vaibhav GuptaCo-founder and Chief Executive · Executive, Boundary (BoundaryML)

“We built an architecture.md file. Instead of using … MD, just pick something that every model can just understand. This file has to be incredibly small, and it can only have things that will not change for months or for years. In our case, it's the layers of the compiler. … So now we have standards so anyone can use whatever they want.”

Vaibhav GuptaCo-founder and Chief Executive · Executive, Boundary (BoundaryML)

“This tool basically visualizes our dependency graph internally with some external dependencies as well, … But what's more interesting is we can go build CLI tools that guarantee that certain invariants can't be broken. And what this does is when Claude builds a new package or adds a dependency that's leaky, we now have CI/CD changing or a simple git commit history that tells us exactly where things break. And by this, we're actually able to make our architecture change. We haven't changed our architecture in the last three or four months.”

Vaibhav GuptaCo-founder and Chief Executive · Executive, Boundary (BoundaryML)

“We do no code reviews. We require every engineer to work on things in parallel, and we have no standardization how people do AI.”

Vaibhav GuptaCo-founder and Chief Executive · Executive, Boundary (BoundaryML)

“I got tired of watching old feature flags pile up… I built an agent for cleaning up stale feature flags. It runs in my local workflow and it doesn't merge any code. The models, the easy part. The judgment in front of it is the whole game. Every flag gets scored the way an experienced reviewer would analyze it. How many modules does it touch? Is it multi-variant? Does it share a component? Then it checks the experiment data. Is the rollout frozen? Is there a sample ratio mismatch? Is there a variant already rolled out to 100 percent?”

Andrew OrobatorSenior Software Engineer · Practitioner, Reddit

“The job changed, and most of us didn't even notice, we're harness engineers now. Our work is the system that helps produce and verify the code. The constraints, the gates, the skills, the verification.”

Andrew OrobatorSenior Software Engineer · Practitioner, Reddit

“It also operates on source code and system configuration only; it does not access or process customer data, payment information, or personally identifiable information.”

BlockPublished by the organization itself (post), no named author

“No longer will they be the sole custodians of building the software. Instead, they will be responsible for building the factories that build the software. With this comes the responsibility of governance, safety, and the ownership of business outcomes.”

FactoryPublished by the organization itself (post), no named author

“No organization starts with a fully autonomous software factory. Autonomy is a maturation process that is gradual and specific to every organization’s readiness and comfort level.”

FactoryPublished by the organization itself (post), no named author

“Well-defined, measurable tasks run with simple Droid agents or skills. Automations coordinate recurring workflows with a shared objective and memory. Remote and persistent execution leverages Droid Computers for long running or local agents. Multi-agent autonomous execution called Missions solve complex tasks over hours or days by decomposing work into parallel tracks to handle.”

FactoryPublished by the organization itself (post), no named author

“I think we found that the state-of-the-art in December was you can probably, run this for about two weeks. By the end of those two weeks, you’d find that, hey, you want to, change the color of a button. Well, it turns out this button is implemented in, 10 different places, and they, have All these different variations, and oh, you forgot one of them, and actually it’s a slightly different color in one spot.”

Walden YanCo-founder and Chief Product Officer · Executive, Cognition

“And I like to define what we've been calling frontier developers inside of Amazon by three behaviors that I've been seeing. One is hands off coding, frontier developers write maybe … of the code that they produce. The rest is agents. The second is that they interact with their agents infrequently. They'll aim to get their coding assistant to run for up to hours at a time without their intervention. And third is that they minimize idle time. These frontier developers tend to run multiple agents in parallel, … through a backlog of tasks.”

Clare LiguoriSenior Principal Engineer, AWS · Practitioner, Amazon

“Currently, we still have humans approving the code, but we see a short path in the near future to a percentage of our code landing automatically, having automatic approvals, right?”

Will BondStaff Engineer, AI Foundations team, Development Platform organization · Practitioner, Uber

“But rather than killing the outer loop, I think that we believe and the industry has just started to really kind of coalesce on this idea that … really expanding the outer loop. Rather than removing humans from the code review process, we are moving their responsibilities up a layer.”

Will BondStaff Engineer, AI Foundations team, Development Platform organization · Practitioner, Uber

“engineers don't trust AI opinions. They just trust clear findings tied to standards that we … agreed upon. … it stopped sounding more like, maybe this is a problem and started sounding more like this violates a standard that we actually care about.”

Rajesh BhatiaSenior Director of Engineering, Cloudflare · Leader, Cloudflare

“Our Network Firewall team audited rampartd using a multi-agent consensus process where every requirement was scored COMPLIANT, PARTIAL, or NON-COMPLIANT with specific violation details and remediation steps reducing what previously required weeks of manual work to a structured, repeatable process.”

CloudflarePublished by the organization itself (post), no named author

“Nearly 45% of issues that start with our AI assistant are now resolved without a human agent while delivering much faster resolution times.”

Brian CheskyCo-founder and Chief Executive Officer · Executive, Airbnb

“nearly 30% of tickets in North America that are English-based are handled by an AI agent”

Brian CheskyCo-founder and Chief Executive Officer · Executive, Airbnb

“Only conversations that were not escalated are evaluated for automated resolutions.”

ZendeskPublished by the organization itself (post), no named author

“Our engineers are orchestrating fully autonomous digital task forces, firing off agents and accomplishing incredible things.”

Google (Alphabet)Published by the organization itself (post), no named author

“The system peaked at ~1,000 commits per hour across 10M tool calls over a period of one week.”

Cursor (Anysphere)Published by the organization itself (post), no named author

“We're now seeing agentic code reviews that look at code changes through a series of different lenses. In the past, it would have been impractical for a cloud infrastructure engineer and a security engineer to review every single code change. With agents, that becomes possible.”

Venkat VenkataramaniVP of Engineering, Applied Infra · Leader, OpenAI

“The number one thing that is changing is that people are starting to use threads for much longer, and this longer usage has been a breakthrough. It's surprising to see the sheer length of time that people spend on a thread – even days! They often set a goal and then have the model crank.”

Andrew AmbrosinoLead, Desktop · Leader, OpenAI