Home›Companies›Plek AI
Plek AI

The AI company brain for PE value creation

Plek is the AI company brain for private equity value creation. Upload an employee census, and Plek maps every person to a taxonomy of 50,000 jobs and 250,000 tasks, down to the time each task takes. It then analyzes spans and layers, org structure, footprint, compensation, functional fragmentation and AI exposure in parallel, and reconciles overlapping levers employee by employee so every dollar is traceable. What takes a performance improvement team 6 to 8 weeks takes Plek under a minute. At a 3,600-person company, Plek found $30M in annual EBITDA levers, matching what a global PE advisory firm found in six weeks. Funds and advisories use Plek from due diligence to exit to run org value creation scenarios, from the platform or from their own AI agents via MCP. Plek grew out of the MIT Work Analytics Lab, founded by CEO Pierre Bouquet during his MIT PhD. Long-term, Plek becomes the Celonis of the workforce: it continuously monitors the org and flags value creation levers as they appear.
Active Founders
Pierre Bouquet is the founder and CEO of Plek, the AI company brain for value creation. He earned his PhD from MIT, where he founded and led the MIT Work Analytics Lab. Before MIT, he founded Biot, an AI logistics layer for healthcare. He holds a BSc and an MSc from EPFL, the Swiss Federal Institute of Technology in Lausanne.
Company Launches
Plek AI: The AI company brain for value creation
See original launch post

TL;DR: Plek AI turns an employee census into a bottom-up value creation plan in under a minute. At a 3,600-person company, it found $30M in annual EBITDA levers, the same number a global PE advisory firm reached in six weeks.

Hi, I'm Pierre. I spent 3 years at MIT building workforce models for my PhD and founded the MIT Work Analytics Lab.

https://youtu.be/WGo6LAYU9A0

The problem

When a PE sponsor runs diligence, acquires a company, or plans a restructuring, it builds a value creation plan. Today, a team of 5 spends 6–8 weeks on it, starting from a blank sheet, and it costs up to 7 figures. Testing another value creation scenario means days more in Excel.

How Plek works: everything in parallel, everything traceable

  1. Map: every employee in the census is mapped to a workforce model of 50,000 jobs in parallel, down to the tasks they do and the time each one takes.
  2. Analyze: all organisation levers run in parallel: spans & layers, org structure vs. peers, footprint and offshoring ROI, compensation outliers, functional duplication, and AI exposure per task.
  3. Deduplicate at the employee level: levers that overlap are reconciled person by person, so savings are never double-counted. Every dollar in the plan traces back to specific roles and tasks.

Run scenarios side by side from the platform or from your own AI agent: all of Plek is available through MCP as a deterministic engine. An optional AI interview agent talks with employees and managers to confirm roles and tasks.

Why Plek is different: the first bottom-up value creation assessment at the task level.

  • ~12,500x more granular than the consulting standard (250,000 tasks vs. ~20 job families).
  • 12x more granular than the best research models (~20,000 tasks).
  • 10,000x faster, from weeks to under a minute.

Results

  • Completed four pilots at companies with 13,000, 3,600, 1000, and 250 employees.
  • Onboarding with a leading PE advisory firm.
  • Deployed at a $2B PE fund.

Where we're going

We're starting with PE because funds need to create value fast and need hard numbers to do it. Long-term, Plek AI becomes the Celonis of the workforce: it continuously monitors the org and flags value creation levers as they appear.

Team

  • Pierre Bouquet, Founder & CEO: MIT PhD in 3 years. Founded and led the MIT Work Analytics Lab, where he modeled the entire US workforce at the task level (published in the International Journal of Production Research). Previously founded Biot and Black Wheel.
  • Melchior Carrier, CTO: EPFL applied mathematics, MIT researcher, founding engineer at Fleetline (acquired in 2026).

Our ask

If you're a PE operating partner, value creation lead, or performance improvement MD, get in touch for a live demo.

Org due diligence is free on dead deals.

👉 Book a demo: cal.com/pierre-bouquet-z3wiko/plek-demo or email pierre@plek.ai

Plek AI
Founded:2026
Batch:Fall 2026
Team Size:2
Status:
Active
Location:San Francisco
Primary Partner:Garry Tan