Before the deal
Know what the data can carry before you buy it.
Every thesis now leans on AI somewhere: churn, pricing, forecasting, collections. Every one assumes the target’s data is fine, and financial and commercial diligence don’t check. We test the AI plays in your thesis against the target’s own records, from data-room exports, and give the deal team a signed verdict, within 5 working days of complete exports.
Starts with a 30-minute call. Commits you to nothing.
The thesis leans on four AI plays. Two have the data today.
Red flags between systems
- High
214 closed-won deals with no matching invoice.
CRM ↔ Billing
Leverability read on the thesis
- Deploy now
Lead scoring
Win or loss history complete for 3 years
What rides on itThesis: new-logo growth of 25% a year
- Deploy now
Pricing optimisation
Prices match the ledger for the core range
What rides on itThesis: 2 points of gross margin
- Needs work
Customer churn prediction
117 customers with no shared ID
What rides on itThesis: net revenue retention to 112%
- Blocked
AI agent on collections
Payment terms live in a spreadsheet
What rides on itThesis: DSO from 58 to 45 days
We use the targets in your plan. We don’t value the plays ourselves.
Two of four plays supported. The other two need named fixes, 6 to 10 weeks, before they count in the plan.
ScopeData-room exports: CRM, billing and HR, 24 months. Not reviewed: live configuration, access logs, product usage. Customer names under clean-team rules. Addressed to the named client.
Checked by hand · Signed off by DataTruth · Ref EX-0902
Illustrative example, not a client.
Real sample
See what a deal team would receive. A full Truth Score Read on 72,837 records in 7 exports from CRM, finance and HR. It scored 59 out of 100, Band C, significant work needed: 823 finance customers with no CRM account, and only 13% of deals reconcile to invoices.
Read the sample memoReal engine output on generated test data, not a client.
Alongside financial and commercial
The third diligence question.
Financial diligence asks whether the earnings are real. Commercial diligence asks whether the market is. Neither asks whether the data under the AI plays in the thesis agrees with itself. Every system looks clean on its own. The truth is in the gaps between them.
From the data room
Exports and consent. A clean team if the deal needs one.
Diligence has its own rules, and the seller sets most of them. We work inside them, and the memo says exactly what they allowed us to see.
-
01
Data-room exports
Standard exports of CRM, billing, ERP and HR, placed in the data room or sent under the deal NDA. We send the field list for the request.
-
02
Seller consent
Customer-level records are used only with the seller’s consent, under NDA and our data processing agreement.
-
03
Clean-team rules, if the deal needs them
Competitively sensitive records stay with named individuals. The buyer sees findings, counts and patterns, not customer lists. How we handle exports
-
04
Scope and limitations
Every memo says what was in the exports, what wasn’t, and what that means for the verdict. A partial read is labelled as one.
What the deal team gets
A verdict, not a data dump.
Engineering findings don’t belong in an investment committee. The memo leads with what the deal team needs to decide, and keeps the evidence underneath.
-
Truth Score™
Can the data be trusted?
A score out of 100 for how far the data can be trusted, on six weighted pillars, with any pillar the exports couldn’t evidence marked not assessed. How the score and its cap work
-
Leverability read
Which thesis plays can run?
Which AI plays the data can run now, after fixes, or not at all; nothing to do with debt. Each play sits beside the thesis target that leans on it. We use the targets in your plan. We don’t value the plays ourselves.
-
Red flags
Where do the systems disagree?
Customers with no billing match, deals with no invoice, accounts owned by leavers: counted and sampled.
-
Effort to close
What would it take?
For each gap, an effort estimate and a test for done, so the fix can be planned into the first 100 days.
-
Scope line
What wasn’t seen?
Exactly which exports and periods were tested, and what live configuration or access logs were out of reach.
-
Verdict
What it means for the plan
Plain language for the deal team, signed off in writing, with a walk-through call before the committee meets.
Timing
Fast enough to inform the decision, not just confirm it afterwards.
- Day 0
Exports and consent in place
The scoping call agrees which systems and which thesis plays; the seller provides exports against our field list.
- Day 2
Matched across systems
Customers, deals and invoices in the exports matched across systems; every red flag checked by hand against sample records before sign-off.
- Day 5
Signed memo and walk-through
Within 5 working days of complete exports, in time for the deal team and before the committee.
- If exclusivity allows
The whole thesis, tested
An AI Investment Review, about 2 weeks: every AI play ranked and a 100-day remediation roadmap written for the new owner.
Calendar time depends mostly on how quickly the data room fills. We send the field list on the first call so nobody guesses.
Before the exit
Find it before the buyer does. Data readiness for the sale.
A year or more before a sale, a Truth Score Read on the systems the buyer’s diligence will open. It is a read on data readiness, not a valuation view, and it makes no promise about price or outcome.
- 01
What their diligence would find
The cross-system checks a buyer’s team could run: customers that don’t match, deals with no invoice, accounts owned by people who have left. Found first, with the records behind them.
- 02
What can be fixed in time
Each gap with an owner, an effort estimate and a test for done, so management can decide what to fix before the data room opens and what to explain.
- 03
The AI story, tested
If the equity story leans on AI, each play is marked deploy now, needs work or blocked, so the story the buyer hears is one the data can carry.
- 04
A memo with its scope attached
Management decides whether the memo goes in the data room. If it does, the scope line goes with it, so nobody reads more into it than it says.
Conflicts and reliance
One side per process. One named client per memo.
One side per process
In any one sale process, we work for one side only.
The named client
The memo is addressed to the named client, and written for them.
Reliance
Anyone else, including lenders, co-investors or a later buyer, can rely on it only with our written agreement.
Independent on the deal as on everything else: we sell no AI, no software and no fixes, take no referral fees from anyone, and our fee is fixed before we start.
Questions
What deal teams ask first.
What a deal team asks before letting us near a data room.
Book the callDo you need access to the target’s systems?
No. We work from exports the seller places in the data room or sends under the deal NDA. No logins and nothing installed.
What if the seller won’t provide customer-level records?
Then the read is narrower, and the memo says exactly how. Aggregated or partly masked exports still show a lot; the scope line records what could and couldn’t be tested, so nobody mistakes a partial read for a full one.
How do you handle competitively sensitive data?
Under the deal NDA, our data processing agreement and clean-team rules, if the deal needs them. Sensitive records stay with named individuals, and the buyer sees findings, counts and patterns rather than customer lists. Our security page sets out how exports are handled.
Can you test the investment thesis itself?
Yes. Send the investment memo or the draft value creation plan. We pull out the AI plays and the growth targets that lean on them, map each to the data it needs, and mark each one deploy now, needs work or blocked.
What happens to the data after the deal?
Raw exports are deleted after the engagement; we keep quality metrics, not records. The full list of sub-processors comes with our data processing agreement, before any data moves. Any transfer outside the EU is covered by Standard Contractual Clauses.
Do you put our data into an AI model?
The Truth Score is rule-based: the same data, scored on the same date, gives the same score, and no AI model decides it. We do use a language model (Anthropic’s Claude, through its commercial API) to read the plan or proposal you send and to help draft the memo’s wording from our findings. A person checks every line before it reaches you. If you’d rather we didn’t use it on your engagement, say so and we do that part by hand.
Can you run diligence on more than one target at once?
We take a limited number of reads at a time, so every red flag is checked by hand. For a portfolio we agree a sequence, one or two companies in parallel, each memo within 5 working days of complete exports.
Does the read follow the company after closing?
It can, if the client agrees. The memo becomes the starting point for the first 100 days, and the company’s leadership can take the AI Investment Review or AI Value Tracking from there. See how funds use the method through the hold.
Will you fix what you find?
No. We never fix what we find. You get a prioritised fix list with an owner, an effort estimate and a test for done on every item, written to hand to your own team or any outside firm.
What does it cost?
A fixed fee, agreed before we start, that never moves with the verdict. It is a small fraction of the proposal it tests; if it wouldn’t be, we’ll tell you on the call. Our first clients get a founding rate in exchange for a private reference call once they’re happy with the work, only with their consent. We never name a client, fund or target publicly without written consent.
Have a target in diligence? Test the thesis on its data.
Tell us the deal timeline and the AI plays in the thesis. We’ll scope the read to fit, on a fixed fee that never moves with the verdict.
Starts with a 30-minute call. Commits you to nothing.