Three material findings your team did not already have, or the engagement is free.
For pilot mandates only. The guarantee does not apply to demos.
What is actually in the room?
A short film: three hundred documents, one weekend, and a clause on page fourteen allowing a customer to terminate on a change of control with thirty days’ notice.
The example in the film is constructed on a fictional target and says so on screen. No client document appears. Your own run produces the same six fields, on your own room.
The room you are reading was assembled by the seller
There is never time to read all of it, and what was never uploaded leaves no trace at all. The price gets set regardless of either.
You are working from a sample
Conventional diligence reads a fraction of the room. The rest is still in there, and it does not stop being true because nobody opened it.
The risk sits between disciplines
A quality-of-earnings read covers earnings. The expensive surprises sit where two specialists each assumed the other was looking.
Why use AI for this?
The short answer, without the hype.
It reads every document
A team under deadline reads a sample. The software reads the whole room, and finds what links one document to another: a clause in a contract that changes a number in the accounts.
Code does the arithmetic
Recalculations across the statements, the LBO model and the sell-side checks run in code, without the language model. A number the model writes that is in none of the quoted passages is held back.
People decide
Your reviewer checks the findings and signs off. Nothing leaves without that.
Whichever side of the table you are on, read the whole room before the price is set.
A leveraged-finance banker and a data engineer
Two people you can reach directly. Wouter structured acquisition and leveraged finance at ING. Daniel builds regulated data pipelines.
Every run goes through the same six steps
Anyone can point an AI tool at a data room. The question is what happens when the model gets something wrong.
- Fabricated or hallucinated citations
- Privacy leaks
- Scope refusals
- Drafts below the reviewer threshold
Every hard-block condition goes to your reviewer, never auto-approved.
Where your documents go, and who can see them
EU hosting by default, zero retention at the model provider, pseudonymisation before analysis, and your own reviewer’s sign-off. Four mechanisms, no certification.
What changes when you sign
You arrive at the table with the whole room behind you.
If you are buying
The price reflects the whole room. Every discipline, reconciled, before the number is set.
If you are selling or brokering the sale
You already know what they’ll find. The buyer’s advisers run this pass anyway. This time you saw it first.
Three steps to get started
The demo runs on your own data room. After it, your team works in the platform itself.
Pick the room
A deal you have open, or one you closed last year. A closed deal exposes nothing in process, and lets you hold what we find against what your team found at the time.
Sign the NDA before a single document moves
Thirty minutes to agree scope and jurisdiction, then the NDA goes out the same day. Access runs under a DPA, in the EU, with zero retention.
Read the findings on a call, then run it yourself
Days later, each finding with its source and location: that call is the demo. After it, your team sets up its own dossiers, invites guests and signs off its own findings.