Poetic

Poetic automates complex, multi-hour enterprise processes in regulated industries by letting operators define workflows in plain English that compile into deterministic, near-tokenless execution. It targets high-stakes financial-services and insurance workflows — fraud investigations, KYC, disputes, and underwriting — where reliability and auditability matter more than open-ended agents.

Poetic, formerly Forge and founded by former Google/Waymo machine-learning engineer Markie Wagner (who previously ran the ML consultancy Delphi Labs), raised a $50M Series A at a $500M valuation on June 10, 2026, led by Kleiner Perkins with participation from Founders Fund, First Harmonic, and OpenAI. Rather than relying on open-ended autonomous agents, Poetic built its own programming language that lets operators define complex workflows in plain English and then encodes that expertise into deterministic, near-tokenless execution — an approach the company says reaches 99% quality on processes enterprises have failed to automate for years, at a fraction of the cost of token-heavy AI agents. The platform ingests training materials such as PDFs, videos, screenshots, and SOPs, learns from expert-operator feedback, self-heals when downstream enterprise systems change, and maintains immutable audit logs with role-based access control. It is deployed at major financial institutions: SoFi reported reaching 99%+ quality on fraud investigations end-to-end within weeks, AIG hit 99%+ accuracy on a multi-hour insurance process, and the platform has processed more than 500,000 disputes and achieved 100% KYC adherence at a top bank. Poetic reached an eight-figure run rate and profitability in 2025 with just four employees, and keeps customer data in-environment with zero data retention.

Category

Automation & Workflows

Pricing

Contact for pricing

Target Market

CIOs, CTOs, Heads of Operations, Financial Services Enterprises

Headquarters

San Francisco, USA

• ✓
  Plain-English workflow programming

  A proprietary programming language lets operators define complex processes in natural language that compiles into precise, repeatable execution.

• ✓
  Deterministic, near-tokenless execution

  Once trained, workflows run deterministically with minimal token usage, cutting cost versus open-ended AI agents.

• ✓
  Learns from expert feedback

  Ingests SOPs, PDFs, videos, and screenshots and refines its decisions from operator feedback to reach 99%+ quality.

• ✓
  Self-healing

  Automatically adapts when enterprise systems or interfaces change, so automations don't break silently.

• ✓
  Immutable audit logs and RBAC

  Maintains tamper-evident audit trails and role-based access controls for regulated environments.

• •
  Fraud investigation auto-decisioning

  Executes end-to-end fraud investigations at 99% quality, as deployed at SoFi.

• •
  Insurance underwriting and multi-hour processes

  Automates demanding multi-hour insurance workflows at 99%+ accuracy, as deployed at AIG.

• •
  KYC compliance and dispute handling

  Achieved 100% KYC adherence at a top bank and has processed more than 500,000 disputes.

Pricing is not publicly disclosed; Poetic sells to large enterprises through a forward-deployed model and positions its near-tokenless deterministic execution as far cheaper to run than token-heavy AI agents.

This page was written from 3 sources, 2 on domains other than poetic.com.

1. 1.poetic.com — poetic.comvendor
2. 2.prnewswire.com — poetic raises 50m series a to automate the worlds most compl
3. 3.kleinerperkins.com — poetic turning enterprise procedures into evolving software