- Poetic raises $50M Sequence A at a $500M valuation led by Kleiner Perkins, with OpenAI, Founders Fund, and First Harmonic additionally investing
- Founder Markie Wagner — a Thiel Fellow and former ML engineer at Google and Waymo — constructed Poetic after years of watching enterprise AI fail the second it met an actual workflow
- The San Francisco startup hit an eight-figure run charge in 2025 with simply 4 workers, posting a 100% pilot-to-production conversion charge at SoFi and AIG
Markie Wagner had a front-row seat to certainly one of enterprise tech’s most persistent failures. As an ML engineer at Google after which Waymo, she watched AI programs that regarded spectacular in managed environments crumble the second they hit the complexity of actual operations — multi-hour processes, hundreds of unwritten guidelines, workflows the place being incorrect 5% of the time just isn’t an possibility.
She left, based Delphi Labs, an ML consultancy deploying AI for among the world’s largest organisations, and saved watching the identical factor occur. The issue was not the fashions. It was the structure. Autonomous brokers, reasoning freely via high-stakes workflows, have been merely not dependable sufficient for the work enterprises truly wanted executed.
So she constructed one thing totally different. That firm is now known as Poetic — previously Forge — and this week it closed a $50M Sequence A at a $500M valuation, led by Kleiner Perkins, with participation from OpenAI, Founders Fund, and First Harmonic.
The repair: software program that learns like AI however runs like code
Poetic’s core departure from the remainder of the enterprise AI market is architectural. Moderately than deploying a mannequin that causes autonomously via a workflow, Poetic constructed a proprietary programming language that lets operations groups describe their most advanced processes in pure language. The platform then converts that logic into deterministic, near-tokenless execution — controllable, auditable, and repeatable at scale.
The goal is the work that has defeated every little thing else: fraud investigations, transaction monitoring, compliance checks, insurance coverage evaluations — multi-hour processes that run hundreds of instances a day and carry near-zero tolerance for error. As experts have noted, the investment opportunity in AI is shifting almost exclusively to vertical, workflow-native solutions — exactly the class Poetic is constructing into. Wagner’s argument is that these workflows are usually not simply technically exhausting. They’re stuffed with institutional information that no person ever wrote down, collected over a long time, and unimaginable to seize in a immediate.
“Proper now in AI, there’s an excessive amount of consideration on fast demos and glossy objects, and never sufficient on outcomes. We constructed a brand new form of software program that learns like AI however runs like code, so the toughest work in your online business lastly will get executed reliably.” — Markie Wagner, CEO, Poetic
What manufacturing truly seems to be like
The numbers Poetic is posting in reside deployments are the type that have a tendency to finish investor debates. At SoFi, the corporate reached 99%+ high quality executing fraud investigations end-to-end in 5 weeks. AIG is a named buyer on advanced, multi-hour insurance coverage processes that beforehand required vital guide effort. The corporate reviews double-digit tens of millions in financial savings for Fortune 500 shoppers — and a 100% pilot-to-production conversion charge throughout each engagement it has run.
That final determine issues greater than it’d seem. The global AI agents market is projected to grow from $7.84 billion in 2025 to $52.62 billion by 2030 — however the enterprise AI graveyard is already stuffed with pilots that impressed in demo situations and quietly stalled earlier than reaching manufacturing. An ideal conversion charge, if it holds as Poetic scales, is the clearest attainable sign that the reliability downside Wagner recognized has truly been solved.
The corporate reached an eight-figure run charge in 2025 with 4 workers — a ratio that claims as a lot about product leverage because it does about capital effectivity.
Why Kleiner Perkins backed it twice
This isn’t Kleiner Perkins’ first wager on Poetic. The firm recently raised $3.5B across two new funds, cementing its place as one of the lively AI traders of the present cycle. It led Poetic’s seed spherical too, and associate Leigh Marie Braswell backed the corporate personally earlier than it had a product to indicate. Doubling down at a nine-figure valuation, with OpenAI and Founders Fund alongside, is an announcement of conviction that goes past a normal follow-on.
“Markie is likely one of the most prescient founders I’ve encountered on AI, and I’ve had a entrance row seat because the starting. What Poetic has constructed is genuinely totally different — a platform that may execute the advanced, high-stakes processes that enormous enterprises truly run, with accuracy that exceeds what human groups can ship.” — Leigh Marie Braswell, Accomplice, Kleiner Perkins
OpenAI’s participation provides a unique dimension. As a strategic investor, its presence alerts that Poetic’s deterministic, low-token execution mannequin is seen as complementary to — reasonably than aggressive with — the broader LLM ecosystem. Wagner herself is a Thiel Fellow, that means Founders Fund’s participation closes a circle that started earlier than Poetic existed.
What the funding is for
Poetic will use the capital to develop its forward-deployed engineering staff, enter new industries past monetary companies, and scale inside its current buyer base. The global intelligent process automation market — the broader class Poetic is competing in — is forecast to achieve $35.8 billion by 2030, with monetary companies, healthcare, and authorities because the three largest verticals. Healthcare authorisations, authorities processing, and provide chain choices are the sorts of workflows the corporate is eyeing subsequent.
Wagner has been direct concerning the bar she is holding herself to: “The enterprise AI panorama is suffering from pilots that by no means made it to manufacturing. For us, we’ve had a 100% pilot-to-production conversion charge. Our know-how works, and we rent one of the best of one of the best — as a result of we’re not right here to run pilots. We’re right here to rework companies.”
The corporate that began with one founder watching AI fail in manufacturing, time and again, now has $50M to show the repair works at scale.
