Poetic: Turning Enterprise Procedures into Evolving Software

Most of the attention in AI applications has gone to categories like writing code, searching enterprise data, and answering customer questions. But there is another category of work, far larger and far less touched by software: the complex, multi-hour processes that keep large companies running. Think insurance underwriting, supply chain operations, financial compliance. Workflows governed by hundreds of edge cases and undocumented rules that live in the heads of experts who have spent years mastering them.

Poetic is building a platform to automate these processes with the level of accuracy that regulated industries actually require. Their core insight is that the hardest part is not the model. It is translating deep human expertise into software that can operate reliably. Poetic built its own programming language that lets operators define workflows in plain English, while the system converts that into deterministic, near-tokenless execution under the hood. At a moment when companies like Uber are openly questioning whether ballooning AI token spend is worth it, Poetic’s architecture is a meaningful differentiator.

I’ve known Markie Wagner, Poetic’s founder, for many years. We first met when she was at Waymo, and she has been one of the most thoughtful and prescient voices on AI I’ve encountered since. Before Poetic, Markie founded Delphi, an ML consultancy that deployed machine learning for some of the world’s most sophisticated organizations. That gave her a firsthand view of what it takes to make AI work in production, not in demos. She waited until the underlying technology was genuinely ready. That patience and conviction are quintessentially Markie.

I was fortunate to back Poetic at the seed stage when Markie was still finishing her last consulting engagement. The company’s progress since has only strengthened my conviction. It was a no-brainer to recently double down and lead this latest $50M Series A round. Poetic reached an eight-figure run rate and became profitable in 2025 with just four people. 

Today, Poetic is deployed at some of the largest financial services organizations, helping automate their most difficult and consequential workflows. In one deployment, a major insurer saw Poetic perform one of its most demanding processes with accuracy that exceeded its own human team.

SoFi CEO Anthony Noto put it well: “In just weeks, Poetic was able to execute many of our fraud processes end-to-end, all while improving quality metrics and the member experience. With auto-decisioning, members get access restored right away, instead of days. As we’ve scaled across the business, we’ve seen 100% process adherence, even in our most complex compliance investigations.” 

We’re excited to deepen our partnership with Markie and the entire Poetic team, and we can’t wait to see what they build next.