What is AI due diligence and why it is becoming a standard
AI due diligence is a systematic assessment of a company's readiness to implement and scale artificial intelligence, conducted by a PE or VC fund before or immediately after closing a transaction.
Traditional technological due diligence assesses IT infrastructure and data security. AI due diligence assesses something different – value potential: does the company have the data, processes, people, and culture that will allow it to leverage AI as a growth driver within the investment horizon?
The scale of change in Western markets is already measurable. According to a 2024 Deloitte study, 65% of PE fund managers are implementing or piloting AI in their investment decision-making process. Bain & Company's Global Private Equity Report confirms that leading funds are already building structured AI assessment protocols – treating them as a standard equivalent to legal and commercial due diligence.
In Poland, this standard is just beginning to take shape. Funds that introduce AI readiness assessment of companies as part of the investment process now gain an informational advantage when evaluating the same targets – before it becomes a market requirement.
The 5D Model – five areas for assessing a company's AI readiness
Assessing a portfolio company's AI readiness requires analyzing five interdependent areas. Omitting any of them provides an incomplete picture – a company may have excellent data and zero cultural readiness, which in practice blocks any AI implementation. We always start the assessment with strategy, not technology.
Diagnosis (Problem Definition)
The starting point for any AI readiness assessment. Can the company name a specific business problem that AI is supposed to solve? Not "we want to implement AI" – but "we lose X zlotys monthly on manual data verification in process Y and we want to shorten it by 60%." A company that cannot formulate a problem in business terms is not ready for AI – regardless of data quality and infrastructure. This is an eliminatory criterion.
Data (Data Readiness)
Data is the foundation of any AI implementation. Data dispersion across systems is not an insurmountable obstacle – AI can aggregate and structure data from multiple sources. The real problem lies elsewhere: whether data actually exists in sufficient volume, whether it is complete, and whether it is reliable. Incomplete or systematically distorted data – for example, selectively collected, depending on who and when entered it – generates AI models that make incorrect decisions with high confidence. This is a risk that the fund must assess before investment, not after.
People and Culture (Adoption)
Technology implemented in an organization with cultural resistance does not work – it only works on paper. The assessment covers two levels: whether the management understands AI as a strategic tool, and whether the management team is ready to change their way of working. This is the most difficult area to assess in standard due diligence and most often determines the success or failure of an implementation.
Governance (Risk & Project Management)
AI without governance is a risk, not an advantage. The assessment includes: whether the company has defined rules for AI risk management – including regulatory, operational, and reputational risk – and whether there is a project management structure for implementation. A company without clearly assigned AI responsibility at the board level is structurally unprepared for scaling, even if the pilot was successful.
Technology (Systems Readiness)
Assessment of the company's technological infrastructure: do operating systems allow integration with AI solutions, what tools are already in use, and is there internal technical competence or a proven implementation partner? A company claiming to use AI should be able to demonstrate a working implementation – not a slide.
AI Maturity Score – how to assess a company on a scale of 1–4
The AI Maturity Score is a synthetic assessment of a company's readiness, resulting from the analysis of five areas of the 5D model. It allows the fund to quickly compare companies, identify gaps requiring intervention, and embed AI in the value creation plan with a realistic timeline.
| Level | Name | Characteristics | What it means for the fund |
|---|---|---|---|
| 1 | Ad Hoc | AI initiatives are informal, experimental, and detached from strategy and processes | High risk. AI implementation first requires building foundations – data, processes, governance |
| 2 | Foundational | First AI actions in selected areas, partial data readiness, nascent governance | Good starting point. Value creation possible within 18–24 months with proper support |
| 3 | Operational | AI actively used in defined processes, measurable results, assigned responsibility | Quick value creation possible. Priority: scaling what works |
| 4 | Strategic | AI embedded in strategic processes, mature data, technology, and governance at the organizational level | Advantage already built into the business model. AI as an argument increasing valuation at exit |
For a PE fund, level 2 is the minimum acceptable for an investment with an active value creation plan. Level 1 does not disqualify a company – but requires a separate budget line and an extended AI return horizon.
How to conduct AI due diligence in practice – a checklist for the fund
The following questions are a starting point for discussions with the company's management. The answers allow for assigning an initial AI Maturity Score and identifying areas requiring in-depth analysis.
Strategy
- What specific business problem do you want to solve with AI – and how much does this problem cost the company today?
- Who on the board sponsors the AI initiative and what is their mandate?
- How does AI fit into the company's development plan for the next three years?
Data
- What data do you collect systematically and for how long?
- Are there situations where data is incomplete or inconsistently entered by different people or departments?
- Which process in the company currently has the best data quality – and why that one?
People
- Have you already carried out any AI project – what worked, what didn't?
- How does the management team react to the prospect of changing their way of working through AI?
- Who in the organization is currently a natural ambassador for AI?
Governance
- Who is responsible for managing AI-related risks – including regulatory risk?
- How do you manage the AI implementation project: methodology, budget, milestones?
- Has the company analyzed the impact of AI Act regulations on its planned implementations?
Technology
- What AI tools are currently used in the company – can you show a working implementation?
- Do the company's operating systems allow integration with external AI solutions?
- Do you have internal technical competencies for AI implementation, or do you rely on external partners?
FAQ
Does a company need to be already using AI to pass AI due diligence positively? No. Level 2 in the AI Maturity Score – meaning initial actions in selected areas, partial data readiness, and nascent governance – is sufficient as a starting point for a fund with an active value creation plan. More important than the number of implemented tools is data quality and the management's readiness to change the way they work.
How does AI due diligence differ from traditional technological due diligence? Traditional technological due diligence assesses IT infrastructure and data security – meaning the current state. AI due diligence assesses value potential – whether the company is capable of generating competitive advantage through AI within the investment horizon and whether this advantage will translate into valuation at exit.
How long does it take to conduct AI due diligence? With external support – 5 to 10 business days. The result is a report with an AI Maturity Score based on the 5D model, a map of red and green flags, and recommendations for the value creation plan.
Do you need support in assessing AI readiness? Contact us.
