# Foaster.ai

> Foaster.ai is a San Francisco startup building an AI-native alternative to management consulting. Its AI agents run 30-45 minute interviews across a company to rebuild an 'operational graph' - a structured map of workflows, handoffs, bottlenecks and information flows - then surface where AI should be deployed first. Human experts review and prioritize the roadmap. Founded in 2026 by Raphael Dabadie and Alexandre Combes, the company joined Y Combinator's Spring 2026 (P26) batch after building two widely shared LLM benchmarks, including the Werewolf Benchmark for social intelligence.

- **Founded:** 2026
- **Headquarters:** San Francisco, California, United States
- **Founders:** Raphael Dabadie (Co-founder & CEO), Alexandre Combes (Co-founder & CTO)
- **Team size:** 2
- **Products:** AI Transformation Assessment, Continuous AI Transformation, Enterprise Deployment, Werewolf Benchmark, The Political Gap Between AIs and Citizens
- **Notable:** Accepted into Y Combinator's Spring 2026 (P26) batch, Werewolf Benchmark shared by Greg Brockman, Clement Delangue and Thomas Wolf, reaching over 300,000 views, Two LLM benchmarks reaching a combined 12M+ views

## Products & services

- **AI Transformation Assessment** — AI agents run 30-45 minute interviews across the organization to capture workflows, tools, handoffs, bottlenecks and undocumented tasks, then rebuild the company's operational graph and identify the highest-impact AI opportunities. Delivered in roughly two weeks.
- **Continuous AI Transformation** — Ongoing monthly support that keeps workflow maps updated, tracks adoption metrics, spots where teams struggle, and delivers personalized upskilling by role.
- **Enterprise Deployment** — Forward-deployed engineers and dedicated transformation teams for larger organizations executing on the roadmap.
- **Werewolf Benchmark** — A research benchmark using the hidden-role social deduction game Werewolf to test LLM social intelligence, deception, bluffing and manipulation resistance. Published as a Hugging Face Space.
- **The Political Gap Between AIs and Citizens** — A research study analyzing whether leading AI models' policy preferences track real electoral outcomes across eight nations, finding systematic ideological clustering.

## Achievements

- Accepted into Y Combinator's Spring 2026 (P26) batch
- Werewolf Benchmark shared by Greg Brockman, Clement Delangue and Thomas Wolf, reaching over 300,000 views
- Two LLM benchmarks reaching a combined 12M+ views
- The Political Gap Between AIs and Citizens study shared by Elon Musk with 12M+ views
- Benchmark found GPT-5 won 96.7% of 210 Werewolf games with a 1,492 Elo rating

## Latest updates

- **2026-03** — Joined Y Combinator's Spring 2026 batch and raised a $130K seed.
- **2026-03** — Launched publicly as 'AI agents that drive your AI transformation' via YC Launch.

## Links

- Website: https://foaster.ai
- LinkedIn: http://www.linkedin.com/company/foaster-ai
- Twitter/X: https://x.com/Foaster_ai

---

Profile page: https://yespress.io/foaster-ai-yc-p26
Published by YesPress — https://yespress.io
Last updated: 2026-07-31