You've heard about artificial intelligence, perhaps about EU regulation. Here's what MB AI does — no jargon — and why it will matter to many businesses in the coming months.
Imagine you want to buy a second-hand car. It looks fine, the seller says it runs perfectly. But you have doubts. So you request a report — a full vehicle history: accidents, service records, real mileage, manufacturer recalls.
That report doesn't change the car. But it changes what you know about it. And what you know changes what you're willing to pay.
MB AI does the same — but for artificial intelligence systems. Not for cars. For the software, algorithms, and AI tools that companies buy, sell, or use.
Except here, it's not an accident we're looking for. It's something more recent and more urgent: does this AI system comply with the new European rules?
In August 2024, the European Union enacted a law on artificial intelligence. It's called the EU AI Act. It's the world's first law specifically regulating AI systems — their design, use, and the responsibilities of those who deploy them.
This law is being phased in. On 2 December 2027, the most important obligations apply. From that date, every AI system classified as "high-risk" — and there are many — must be fully compliant.
What is a "high-risk" AI system? Examples include:
| Type of system | Concrete example | Status |
|---|---|---|
| AI-assisted recruitment tool | Software that screens CVs and scores candidates | High risk |
| Automated credit scoring | Banking algorithm that decides whether to grant a loan | High risk |
| Industrial predictive maintenance | AI monitoring equipment in a factory | High risk |
| Medical diagnostic tool | Software that helps a doctor read a scan | High risk |
| Customer service chatbot | Virtual assistant on an e-commerce site | Minimal risk |
| Anti-spam filter | Automatic email sorting | Minimal risk |
If your company uses or sells a system in the first category, you're affected. And most companies don't know it yet.
Since this law came into force, many companies have become aware of the issue. Some consulted their lawyers. Others approached the major consulting firms — the leading audit firms.
The problem? Lawyers read legal texts but don't quantify risk. The major firms produce excellent reports — but in 4 to 6 months, for €150,000 to €500,000. That's fine if you have the time and budget.
But in real life, decisions are not made over 6 months.
An investment fund wants to acquire a startup that does AI-driven recruitment. The deal must close in 3 weeks. Nobody on the team knows whether the software complies with European law. Without that information, they sign — or don't sign — without really knowing why.
That is exactly the problem MB AI solves.
MB AI analyses an AI system and produces a score. That score is called the D7™. It runs from 0 to 100. It measures 7 dimensions: regulatory exposure, technical maturity, data governance, AI governance, commercial defensibility, financial value adjustability, and operational risk.
The verdict: GREEN, AMBER, or RED.
And it takes 48 hours. Not 6 months.
MB AI uses artificial intelligence to automatically collect and analyse all publicly available information on the target company — its website, press coverage, patent filings, official registries, code repositories. An expert analyst then validates the result.
What leading audit firms do in 6 months with 20 consultants, MB AI does in 48 hours — because the repetitive, documentary work is automated.
A structured memo. For each dimension: a score, a justification, identified risks, and a phased remediation plan. Each item is tagged by confidence level: PROVEN (direct source), INFERRED (logical deduction), or ABSENT (missing information).
This memo is directly usable in a negotiation, an investment file, or an insurance underwriting process.
When a fund or a corporate acquires an AI startup, it inherits all its regulatory problems. The D7™ score identifies them before signing — not after. A 10% haircut on a €100M asset is €10M of value. The memo costs €15,000.
W&I insurance policies increasingly cover risks related to corporate transactions. But to assess AI risk, you need a reliable score. The D7™ becomes the reference document for underwriters at AXA XL, Howden, and Tokio Marine.
An investor who receives a GREEN D7™ score with a funding deck has proof that the AI system is compliant. That's a concrete advantage over competing decks.
Bpifrance and the European Investment Bank finance hundreds of AI startups. They need to know whether those startups comply with the new rules — not month by month, but quarterly, across their entire portfolio.
MB AI is not a generalist consulting firm that added "AI Act" to its website. It's a tool built specifically for this problem, by a team that comes from where the rules are made — not from where they are commented on.
The D7™ methodology is proprietary. It synthesises the leading audit-grade risk frameworks, adapted to the EU AI Act — reading each asset under every obligation to surface exposures before a buyer or insurer does.
MB AI does not replace lawyers. Lawyers draft contracts, defend clients, interpret legal nuances. MB AI identifies and quantifies risks before lawyers need to intervene.
MB AI does not replace major consulting firms for large-scale organisational transformations either. If a company needs to completely overhaul its AI governance over 18 months, that's a leading audit firms project. MB AI is the tool that says whether that project is necessary — and how urgent it is.
MB AI is the technical inspection report for an AI system — produced in 48 hours, defensible before a judge or an investor, at a fraction of the cost of a traditional audit.
Because 2 December 2027 is not an abstract date. European authorities have already opened their first investigations. The CNIL in France is designated as the competent national authority. First sanctions will follow.
And once a fine lands, it is too late to negotiate the price of an asset or structure an escrow. The risk has already materialised.
Every week without a D7™ score on an AI asset is a week of unquantified risk. Not hypothetical. Real. Measurable. Avoidable.
A D7™ score in 24 hours.
Less expensive than one hour of legal advice. More useful than a 6-month report.