Assay by KOBECO · technology due diligence platform

Know what’s real before you sign.

Assay is KOBECO’s technology due diligence platform, led by Wouter Denayer, former CTO of IBM Belgium. It reads the data room and the code, answers the due diligence questions, and shows you where the two disagree.

AI agents and measurement toolsFor private equity, venture capital and corporate acquirers

Assay · AI checkWrapper or moat?
From peopleOur AI is our moat: proprietary models, trained on our own data.
Measured in the codecalls to one third-party model = 100%No training or fine-tuning code found.
Thin wrapperThe claim does not hold.Fictional example
The platform

The documents, the code, and the people.

Every KOBECO due diligence runs on Assay: AI agents read the data room and the code, measurement tools measure the code. The people are where I come in.

Assay reads

The documentation

How the company describes itself: architecture, security, roadmaps.

Assay reads

The code

How the product is really built: quality, security, licences, who builds it.

+ Wouter

The people

Missing pieces, open questions, and how the team works. That only comes out face to face.

How it works

From data room to decision in four steps.

  1. ConnectUpload the data room to Assay and give it read-only access to the relevant source code repositories.
  2. AnswerAssay answers the questions in the due diligence framework.
  3. Cross-checkWhere a document and the code disagree, Assay records it and raises it as a finding.
  4. DecideYou get a first analysis before any interview. Once I validate it, you get my judgment and a report I sign.
What only Assay shows you

Where the story and the code disagree.

Every contradiction shows both sides.

What the data room saysWhat the code showsStatus
Multi-tenant architecture, fully isolated per clientRead from documentsShared database, tenant filter applied in 3 of 5 servicesMeasured in the code · billing-api, core-apiVerified
No copyleft licences in the productRead from documents2 AGPL-3.0 packages shipped in the customer-facing buildMeasured in the code · web-appVerified
Automated test coverage is above 80% across the platformRead from documentstest coverage = 31%Measured in the code · 6 repositoriesVerified

Examples are fictional.

Evidence on every answer

Every answer shows how it was produced.

Click any answer and see what it rests on.

Measured in the codeRead from documentsFrom people

AI due diligence

Real capability, or a wrapper?

Almost every target now says it has AI. Assay measures what sits behind the claim: whose models it uses, what it owns, and what happens when the provider changes its price or goes down.

  • Model vendors. Which providers the product calls, and how concentrated the calls are.
  • Switching cost. Whether changing provider is a setting or a rewrite.
  • Evaluation. Whether the team tests model output before it ships.
  • Own models and data. Code that trains or adapts models, and proprietary data that feeds them: the usual source of a real advantage.
  • Fallback. What the product does when the provider fails.

No AI is not a weakness: a product without AI gets “not applicable”, never a lower score. If the code cannot be measured, the AI questions stay open.

Sample AI profileFictional target
Model callsProvider A 82% · Provider B 18%
Switching providerpossible, one layer
Evaluationruns in CI
Training codenone found
Own data pipelines2, feed retrieval
Fallback on outagenone
Thin moatThe data is theirs. The model is not. One provider carries 82% of calls with no fallback: a price rise there lands directly on gross margin.
Where Assay fits

Assay is the first level of every due diligence.

Every due diligence I run starts with Assay. Then I validate what it found, interview the people face to face, and stay after the deal if you want.

1 · always, automated

Assay report

The data room and all the code, read and cross-checked. A first analysis, fast.

2 · + Wouter

Validated due diligence

Findings validated, interviews face to face, gaps closed. A report I sign.

3 · + Wouter, after the deal

From report to results

The improvement plan, the first 100 days, a seat at the board table.

See how the levels build

Who is behind Assay

Wouter Denayer. Twenty years at IBM: IT architect, then CTO for Belgium and Luxembourg, and the global technology strategy team. Four patents in artificial intelligence. Clients include Summa Equity and Fortino Capital.

You work with me directly, from the first call to the board meeting. I lead every assignment and sign every report. More about me

Recent work

Life-sciences AI.PE · buy side · US

Sold as a platform; the code showed much narrower scope. Valuation cut.

$400M life-sciences merger.PE · buy side · US/Belgium

Three due diligences, then helped merge 200 engineers into one team.

Mortgage platform.VC · sell side · Belgium

Core sound, one pillar obsolete. Rewrite it and fix security before the exit.

“Wouter consistently delivers the technical breadth and depth we need to invest with confidence.”

Hans Cobben · Partner, Summa Equity

“Wouter is an amazingly technical consultant with a breadth of experience in software engineering, web applications, large scale data management, ontology management, security, and artificial intelligence.”

Rakesh Nagarajan · Chief Medical Officer, Velsera

See Assay on a deal you are looking at.

A 30-minute call is enough to see if your deal fits.

Book a 30-minute call