AI readiness assessment: what to look for before acquisition
Most acquisition theses now include some version of "AI opportunity" in the value creation plan. The problem is that very few deal teams have a structured way to assess whether that opportunity is real, accessible or priced correctly.
The gap shows up at portfolio review, twelve months post-close, when the AI workstreams that looked straightforward in the CIM have stalled. The data is messier than expected. The technical team is thinner than it appeared. The integrations that were supposed to take three months are now in month nine.
This is not a technology problem. It is a diligence problem.
Most operating partners are rigorous about commercial, financial and legal risk. AI and data readiness still gets treated as a footnote - a few questions to the CTO, a scan of the tech stack slide in the data room. That is not sufficient for businesses where data infrastructure and AI capability are either core to the value creation thesis or central to operational risk.
This article sets out a practical framework for assessing AI readiness during acquisition diligence - what to look for, what the warning signs are and how to avoid buying a problem you did not see coming.
Why AI readiness belongs in every deal, not just tech deals
The instinct to treat AI diligence as relevant only for software or technology businesses is understandable but increasingly wrong.
Consider a mid-market distribution business turning over £600m. The investment thesis is operational efficiency: route optimisation, demand forecasting, warehouse automation. All three workstreams depend on clean, accessible, historicised operational data. If that data sits in a legacy ERP that has been customised beyond recognition, or if the forecasting logic lives in spreadsheets owned by one operations director, the AI-driven efficiency gains do not materialise on the timeline the model assumes.
The same applies to consumer businesses where personalisation is the growth lever, to B2B services businesses where AI-assisted pricing or contract analytics are part of the value creation plan, and to any business where an acquirer intends to consolidate portfolio data across multiple entities.
The question is not "is this a tech company?" The question is: "does the value creation plan depend on data and AI capability that may not exist?"
If the answer is yes, formal AI readiness assessment is not optional.
The four dimensions of AI readiness
A rigorous AI readiness assessment covers four areas. Each one can independently break a value creation thesis.
1. Data infrastructure and quality
This is the foundation. AI systems do not create good data - they depend on it. Assess whether core operational, commercial and customer data is structured, accessible and historicised to a sufficient depth. Ask specifically: where does the data live, who owns it, can it be extracted without the current ERP or CRM, and has it been validated against a third-party source at any point?
Red flags include data that exists only in on-premise systems with no API access, significant gaps in historical records (particularly around customer transactions or product-level margin), and reliance on manual reconciliation processes to produce management information.
2. Technical team capability
Distinguish between teams that consume AI tools and teams that can build and deploy AI systems. A business with three data analysts using off-the-shelf BI software is not the same as a business with an in-house data engineering function and a machine learning capability. Both can be fine - but they need different investment plans and different timelines.
The risk to flag is a business that presents a capable technical function but where one or two individuals carry the institutional knowledge. If those individuals leave post-close, the AI roadmap leaves with them. Key-person risk in data functions is just as material as key-person risk in sales or operations.
3. AI and data governance
This is the area most frequently skipped. Assess whether the business has documented data ownership, data retention policies and any form of model governance. For regulated sectors - financial services, healthcare, consumer credit - this is non-negotiable. For unregulated businesses, it still matters: the absence of governance creates both legal exposure (particularly under UK GDPR) and operational fragility.
Ask whether the business has deployed any AI systems and, if so, whether those systems have been documented, tested and monitored for drift or failure. Undocumented AI in production is a liability, not an asset.
4. Integration and scalability readiness
Assess how easily the business's data environment can connect to external systems, absorb new data sources or scale to support new AI workstreams. This matters particularly in buy-and-build strategies where the acquirer intends to integrate multiple entities into a consolidated data platform.
A common scenario: a PE-backed business acquires a second entity with the intention of running a shared analytics platform within 18 months. Two years later they are still running parallel systems because the legacy architecture of one entity cannot be migrated without a full ERP replacement. That outcome was visible in diligence - it simply was not looked for.
What the data room will not tell you
Standard data room materials - management accounts, customer data summaries, technology stack documentation - are produced to support a transaction. They reflect how the business wants to be seen.
The information that matters for AI readiness is rarely in the data room. It has to be elicited directly.
Request access to sample data extracts and test them. Do not accept a data dictionary as a substitute for the data itself. Ask the technical team to walk through how a specific analytical output - a customer churn report, a demand forecast, a margin analysis by SKU - is actually produced. The gap between what the slide says and what the process actually involves is usually instructive.
In management interviews, ask open questions about recent AI or data initiatives that did not go as planned. Every business of meaningful scale has them. How the team describes those failures tells you far more about technical maturity than a list of tools in use.
Where the diligence timeline allows, a structured technical deep-dive with an independent third party is worth the cost. A focused AI readiness diagnostic can be completed in two to three weeks and will surface issues that a generalist due diligence workstream will not catch.
Translating findings into deal terms
AI readiness findings should feed directly into three aspects of deal structuring.
Valuation adjustment. If the value creation plan assumes AI-driven EBITDA improvement and the data infrastructure cannot support that roadmap without material investment, the cost of that investment should be reflected in the entry price or modelled explicitly in the returns case.
Post-close investment planning. AI readiness findings should translate into a specific, costed 100-day plan for data and AI infrastructure. Vague commitments to "build out the data capability" post-close are not a plan. Name the systems, the roles that need to be hired or contracted, and the milestones.
Warranty and indemnity scope. Where the diligence has identified undocumented AI systems, data compliance gaps or reliance on third-party data sources without clear licensing, ensure W&I coverage is adequate and that specific representations have been obtained from the vendor.
Deal teams that treat AI readiness as a binary pass/fail miss the point. The output should be a risk-adjusted view of the value creation timeline, with specific line items for what needs to be true about the data and AI environment for the thesis to hold.
The cost of getting this wrong
The businesses that stall on AI post-acquisition do not fail dramatically. They fail quietly - in the form of delayed initiatives, consultant engagements that go nowhere and management bandwidth consumed by data problems rather than growth.
By the time those costs are visible, the deal is closed and the price has been paid.
Operating partners who build AI readiness assessment into diligence as a standard workstream - not an afterthought - avoid that outcome. They also arrive at close with a clearer mandate for the post-close technology agenda, which accelerates execution.
If you are approaching a transaction where AI or data capability forms part of the investment thesis, a structured pre-close diagnostic is the right first step. Rodan's diligence-focused AI readiness assessments are scoped for deal timelines and deliver a clear, commercially framed view of what you are buying, what it will cost to develop and what the realistic value creation trajectory looks like.
Speak to our team before the process moves to exclusivity - that is when there is still room to act on what you find.
