fractional AI for European tech companies

Everyone's adopting AI.
Nobody's transforming with it.

I help Founders and CEOs deliver on AI in weeks — by showing up and building it with you.

The problem · 01

Your team uses AI.
But nothing feels different.

AI-native companies aren't 20% more productive. They're unlocking 100× more capabilities.

Most teams are stuck at Level 1 or 2 — AI as a thought partner or personal assistant. Useful, but individual. The org hasn't changed.

Real value lives at Level 3 (AI as a teammate that runs actual workflows) and Level 4 (AI as the operating system running the org). I've studied how 18 of the best AI-native companies got there. The pattern is specific, concrete, and crossable in weeks.

Most teams just don't know what the next step is.

Sound familiar?
  • "We rolled out Copilot. Adoption is 30%."
  • "Every team built its own GPT. None of them talk."
  • "We can't point to a single workflow AI has changed."
  • "Exec team discusses AI more than they use it."
  • "The AI committee meets monthly."
  • "We keep waiting for the governance doc."
The model · 02

Four levels.
Most of your team is on the left.

I mapped the AI journey across 18 companies, Anthropic to Zapier: thought partner → assistant → teammate → the operating system. Levels 1–3 are steps. Level 4 is a different kind of company. The path there runs through each level, one function at a time.

Level 01
Thought partner
Individuals ask AI for ideas and answers. One-off prompts, no workflow change.
Tool: ChatGPT tab
Level 02
Assistant
Individuals complete tasks faster. Time saved, but the org still runs the same way.
Tool: Copilot · 30% adoption
Start here
Level 03
Teammate
Agents run inside team workflows. Real work gets automated, not just accelerated.
Tool: Team agents · workflows
The destination
Level 04
The operating system
AI is the system running the org. Company / Team / Individual / Project — all connected, always current.
Tool: AI Chief-of-Staff OS
The path: I accelerate teams from L1/L2 to L3 in weeks. Then function by function, L3 installs accumulate into L4 — a company-wide operating system. You don't jump to L4. You build it, one team at a time.
Who I am · 03

I've operated at Level 3–4 for years.

I'm the ex-CEO of a YC company. My team went on to found Lovable, Legora, and Tandem Health. I run my own work on agents by default. This is how I worked before I started selling it.

I'm not a strategist who read a blog post and made a deck. I'm an operator who's been living this for years, and knows exactly what it takes to get a team there.

Lovable
Legora
Tandem Health
+ Y Combinator network
Background
Ex-Y Combinator CEO · operator
Default operating level
Level 3–4 · for years
Live clients
2 live · embedded 6 months each
Source library
18 AI-native company playbooks
Not
An agency. A strategy deck.
Engagements · 04

The audit is the door.
Everything starts there.

2 slots open. If this reads like your company, reach out.

Proof · 05

Two live clients.
Eighteen company playbooks.

Live clients · 2026

2 clients live.

Embedded Chief of AI · 6 months · Series B/C EU tech
  • First agent live in days; adopted by the entire team within weeks
  • Strategy decisions that took hours now take under 10 minutes
  • Phased rollout: primary user → full project and team. Measurable adoption at each step.
The source library · what I install from

18 company playbooks.

Agents, workflows, and operating patterns from the best AI-native companies
  • Anthropic · OpenAI · Google · Perplexity · Block · Shopify · Klarna
  • Cursor · Cognition · Harvey · Sierra · Zapier · Midjourney · Notion · xAI
  • Ramp · Coinbase · Box — patterns that transfer to non-AI-native orgs

2 slots open.
The audit is where it starts.

If this reads like your company, reach out.

Book the audit →