We teach your deal team to be independent with AI
Know how the machine works, not just which buttons to press. We set your AI strategy, then either train your deal team to build its own agents or build them for you, on your data and templates.
Discuss your AI strategyOur team comes from
Collaborations with
What AI can do for your deal team
Workflows we build for private equity, M&A and transaction services teams.
Build-up screening
- 10,000+ companies sourced in a sector
- Enriched from Pappers, LinkedIn and websites
- Filtered to ~400 high-fit targets
Market mapping
- Every player in a market
- Expertise, ownership and revenue
- Analysis table and full report
Notes, CRM and decks
- Meeting notes into your CRM
- Excel analyses updated automatically
- PowerPoint materials refreshed
One-pagers and pitchbooks
- Any company, on demand
- Your exact template and layout
- Standard teaser and pitch pages
Valuation
- Full DCF model and report
- Trading and transaction comps
- Football field and sensitivities
Buyer outreach
- Acquirers ranked by fit
- Personalised teaser emails
- Replies and follow-ups tracked
Data rooms and diligence
- Document triage and summaries
- Answers cited to the source
- DDQ pre-filled from past answers
Data production
- General ledger to standard FEC
- Lease debt schedules
- People database from payroll
Deliverable review
- Consistency checks on VDD/BDD
- IC materials and memos
- Slide headers in your house style
How we work
Map
We list your team’s tasks, pick the use cases worth doing and set your AI strategy. Scoping note in 3 working days.
Train
Learn to build on your own. Hands-on workshops on your live deals and documents, until your team builds and extends its own tools.
Build
Or we build with AI for you. We make AI agents, automations and skills on your files and templates, inside your confidentiality perimeter.
Hands-on and thoughtful
Designing an AI agent starts with workshops where your team maps the current process and agrees on the target workflow. Here is what we have learned doing it, and how we work around it.
The whiteboard is not the real process
Workshops map the official workflow, but on the ground a large share of the work is exception handling: the edge cases, the workarounds, the “it depends”. We map those too, because that is where the agent will fail first.
Calibrating what the agent needs
Some teams expect great output without giving enough context or evaluation tests. Others underestimate the model and keep more humans in the loop than needed. We set the right level of instructions, tests and review together.
Prototype fast, then iterate
We put a working prototype in front of your team as early as possible, and say upfront that it takes feedback, iteration and maintenance to reach the quality of a junior analyst.
A central knowledge base starts with how people write
Agents work best on plain text (ideally Markdown) they can read directly or search. Your knowledge sits in Confluence, SharePoint, Word, PowerPoint, PDFs and personal notes, the slowest formats for an agent to use. We help you build one central knowledge base that serves both people and agents, starting with a simple habit: no more PowerPoint for internal meetings.
Past use cases & results
A deal CRM connecting every execution file, with dedicated agents for IMs, buyer prospecting, process tracking and due diligence.
Trade reconciliation automated. Backlog cleared to T-1.
Due diligence and reporting run on infrastructure the firm controls. Proprietary data never leaves.
The team
Emilie Carrel
Former investment banker. Professor in AI x Finance at Sciences Po (teaching & building AI agents)
Karim Michel Nseir
Former BCG consultant. Professor in AI x Finance at Sciences Po (teaching & building AI agents)
Benjamin Mardel
Anton Riahi
Ready to be independent with AI?
Tell us the workflow your team wants to own. 30-minute call, scoping note in 3 working days.
Book a call