Poetic, a San Francisco startup, has raised $50 million in a Series A round to expand its software platform for automating complex back-office processes at large enterprises. The company is led by CEO Markie Wagner, and describes what it has built as software that "learns like AI but runs like code."
Banks, insurance companies, and healthcare providers still rely heavily on manual work and decades-old software for many of their core operations. General-purpose software has improved parts of these workflows, but it was never designed for the kind of messy, rule-heavy, high-stakes processes that define these industries. AI has been widely pitched as the answer, yet most deployments have fallen short. According to MIT's NANDA initiative, 95% of enterprise generative AI pilots fail to produce measurable financial results, with most never making it to full production. Poetic is built specifically around that problem.
Kleiner Perkins led the round, with Founders Fund, First Harmonic, and OpenAI also participating. The raise values Poetic at $500 million.
Both traditional software and newer AI agents run into the same barrier when applied to complex enterprise work. Standard code is too brittle to keep up with constantly changing workflows. AI agents, while more flexible, can be unpredictable in ways that regulated industries simply cannot tolerate. The sectors where this matters most, financial services, insurance, and healthcare, operate under strict compliance requirements and have little room for error.
Poetic's reported results offer a concrete illustration. At SoFi, the US digital bank, the platform reached 99% or higher accuracy running fraud investigations end-to-end, completing the deployment in just five weeks. At AIG, one of the largest insurance companies in the world, it hit the same accuracy threshold on a multi-hour process that previously required significant manual work. These are not simple tasks. Fraud investigations and insurance workflows carry real regulatory weight, and errors in either area can have serious consequences for both the company and its customers.
"Markie is one of the most prescient founders I've encountered on AI, and I've had a front row seat since the beginning. What Poetic has built is genuinely different - a platform that can execute the complex, high-stakes processes that large enterprises actually run, with accuracy that exceeds what human teams can deliver. The fact that they were able to automate at the largest companies with the highest requirements is a reflection of how deeply the product works. And they've done so with one of the strongest teams - leaders from Palantir, UiPath, Ramp, Scale, Retool, and many more." — Leigh Marie Braswell, Partner, Kleiner Perkins
The new funding will go toward hiring and expanding Poetic's forward-deployed team, the engineers and specialists who work on-site with clients to design and implement automations. The company also plans to move into industries beyond financial services, with healthcare and insurance identified as the next priorities.
Speed of deployment has been a notable feature of Poetic's early work. Reaching production quality in five weeks is well outside what most enterprise AI projects manage. Most get stuck in a long pilot phase and never move beyond it.
"The enterprise AI landscape is littered with pilots that never made it to production. For us, we've had a 100% pilot to production conversion rate. Our technology works, and we hire the best of the best - because we're not here to run pilots. We're here to transform businesses." — Markie Wagner, CEO and Founder, Poetic
Wagner previously worked as a machine learning engineer at Google and Waymo, and ran an AI consultancy called Delphi Labs before founding this company, originally under the name Forge. The rebranding to Poetic came later.
The core of the product is a programming language Poetic built specifically for enterprise automation. Rather than instructing an AI agent to figure out a workflow on its own, operators describe the process in plain language. The system then translates that into a set of structured, step-by-step instructions that behave more like traditional code than open-ended AI. The practical effect is that the system runs consistently, without the unpredictable behavior that makes AI agents difficult to trust in regulated environments. It can also ingest a company's existing procedures, training materials, and expert feedback, and repair itself when the underlying software it interacts with changes.
Processes that run thousands of times per day, depend on rules that were never formally written down, and require near-perfect accuracy are the company's target. Poetic reached an eight-figure annual revenue run rate in 2025 with a team of just four employees. Current clients include SoFi, AIG, and Chime. The company says every pilot it has run has converted to a full production deployment.
Kleiner Perkins led the Series A. The firm has a long history of investing in enterprise software companies. Founders Fund, First Harmonic, and OpenAI also joined the round. OpenAI's involvement is worth noting: Poetic's platform is partly a response to the limitations of standard large language model integrations in production settings, which makes OpenAI an unusual but not illogical backer.
