Every deal team we speak to is wrestling with the same question: how do you value AI in a target company? The seller's deck promises transformational AI capability. The diligence data room contains a handful of model cards, a Databricks invoice, and a LinkedIn list of data scientists. Somewhere between those two artefacts sits the real answer — and the gap between them is where deal value is lost, over-paid, or silently written down twelve months after close.
AI due diligence is not a technology exercise. It is an intangible asset exercise. The engineers, the training data, the fine-tuned models, the prompt libraries, the MLOps pipelines — these are capital assets, and they behave like capital assets. They have a cost of creation, a service life, a risk profile, and a defensible market value. The Opagio growth platform was built to measure all of that.
In a standard commercial due diligence, AI sits awkwardly across three workstreams and is properly covered by none of them. Technology DD looks at the stack; commercial DD looks at revenue impact; financial DD looks at capitalised costs. None of them produce a number that can be carried onto the opening balance sheet or defended to a buyer at exit.
The result is predictable. Either the acquirer pays a premium for AI narrative that evaporates post-close, or the seller gives away real AI capability for free because it never surfaced in the CIM. Both outcomes are measurement failures, not judgement failures.
AI value is not a line item in the P&L. It is a stock of intangible capital that produces a flow of economic benefit. Diligence needs to measure the stock, not just the flow.
Opagio's taxonomy maps AI capability into four of The Opagio 12 Value Drivers. Every AI due diligence on the platform works through these four categories in sequence.
| Category | What it includes | Typical valuation method |
|---|---|---|
| Technology capital | Proprietary models, fine-tuned weights, prompt libraries, inference infrastructure, MLOps pipelines | Relief from Royalty, Cost approach |
| Data capital | Training datasets, labelled corpora, embedding stores, feedback loops, data rights | MPEEM, Cost-to-recreate |
| Organisational capital | AI governance, model risk management, deployment playbooks, institutional know-how | With & Without, Cost approach |
| Human capital | AI/ML engineers, data scientists, prompt engineers, AI product managers | Replacement cost, Earnings differential |
Every asset in the target company's AI stack sits in one of these four buckets. The platform's job is to make sure none of them are missed, and that each is valued with a method that an auditor, a seller, and a future buyer will all accept.
The workflow is the same whether you are a PE firm looking at a £40m SaaS add-on or an M&A advisor preparing a sell-side CIM. It runs in five stages.
The target completes the Intangible Asset Questionnaire, which is pre-configured with AI-specific branches. Instead of asking "do you have AI?" it asks the questions that actually surface value: how many proprietary models are deployed in production, what training data do you own outright, which customer feedback loops improve the models, and what proportion of the engineering team can rebuild the stack from scratch if the lead departs tomorrow.
The output is a complete asset register, not a marketing claim. Every asset is tagged with its category, its estimated service life, and the evidence the target has provided to support it.
Self-reported data is necessary but not sufficient. The platform cross-references every claim against three independent signals: capitalised cost recorded under IAS 38 (or ASC 350 for US targets), payroll evidence for the AI team, and infrastructure spend extracted directly from Xero, QuickBooks, or the target's cloud billing console. When those three signals disagree, the platform flags it. That flag is the diligence team's cue to ask harder questions.
A recent sell-side diligence surfaced a target claiming £1.2m of proprietary AI IP. Payroll showed two AI engineers, cloud spend showed £18k of inference cost, and capitalised R&D under IAS 38 was zero. The flag saved the acquirer from a multi-million pound over-pay.
Each asset gets the valuation method that fits its economic profile. Proprietary models with clear revenue attribution go through MPEEM. Brand and patent-adjacent AI IP goes through Relief from Royalty. Pre-revenue technology capital goes through the cost approach. Organisational capital with step-change revenue impact goes through With & Without.
The platform does the method selection automatically based on the asset's properties, but every selection is overridable and auditable. The full method rationale appears in the final report.
AI assets have high valuation uncertainty. A model that looks transformational today may be a commodity in two years. The platform runs three scenarios for every AI asset: base case, downside (model commoditisation, data rights challenge, key-person departure), and upside (successful expansion into adjacent use cases). The range, not the point estimate, is what goes into the deal model.
This is where AI due diligence diverges most sharply from traditional intangible asset valuation. The service-life assumptions for AI models are significantly shorter and more uncertain than for a brand or a customer list. The platform's sensitivity module is built specifically to expose that risk rather than hide it.
The output is a report in the format your auditor already recognises. Every number traces back to its source data, every method is documented, every assumption is exposed, and every scenario is priced. The full audit trail is stored in the platform and can be handed to the buyer's diligence team at exit — which matters, because the number you put on the balance sheet today is the number you will have to defend in two, three, or five years.
Three things change when AI due diligence is run through a structured intangible asset platform rather than a Word document and a spreadsheet.
First, the buyer stops paying for narrative. If the AI capability cannot be evidenced, categorised, and valued against a defensible method, it is not in the price. This alone recovers deal value that would otherwise be forfeit to seller optimism.
Second, the seller stops leaving value on the table. Real AI capability that is buried inside "technology" or "R&D" gets surfaced, named, and priced. In three recent sell-side engagements we ran this framework against, the headline valuation moved up by between 8% and 17% purely because assets that already existed were made visible.
Third, post-close value creation becomes a measurable workstream rather than a hope. The same asset register that priced the deal becomes the portfolio dashboard that tracks whether the thesis is actually playing out. If the AI-driven gross margin uplift was valued at £4m of excess earnings, the operating partner now has the line item to monitor.
The same framework that protects you in diligence becomes the value-creation plan post-close and the exit positioning three years later. Diligence, value creation, and exit are the same conversation — the platform just makes them the same data.
AI due diligence on the Opagio platform is actively used by four buyer personas.
PE firms running bolt-ons and platform acquisitions where AI capability is part of the investment thesis. The framework tells you whether the AI is real, pre-close. The portfolio dashboard tells you whether it is scaling, post-close.
M&A advisors preparing sell-side CIMs for AI-native or AI-enabled businesses. The framework surfaces value the financial accounts do not show and gives the buyer a defensible number to underwrite against.
Corporate acquirers acquiring AI capability as a strategic capability rather than a standalone revenue stream. The framework translates "we bought an AI company" into capitalised intangible assets that survive the integration.
Lenders and credit committees underwriting AI-heavy borrowers. The framework produces an asset register defensible enough to support IP-backed lending decisions — which is the fastest-growing use case we are seeing in 2026.
A mid-market SaaS target self-reported AI capability worth £6m in the management presentation. The platform ran the framework:
| Asset | Category | Method | Base case | Downside | Upside |
|---|---|---|---|---|---|
| Proprietary ranking model | Technology capital | MPEEM | £2.4m | £1.1m | £3.8m |
| Labelled training corpus | Data capital | Cost-to-recreate | £0.9m | £0.7m | £1.4m |
| Model deployment playbook | Organisational capital | With & Without | £0.7m | £0.3m | £1.2m |
| AI engineering team (4 FTE) | Human capital | Replacement cost | £1.1m | £0.9m | £1.5m |
| Total AI intangible value | £5.1m | £3.0m | £7.9m |
The acquirer entered negotiations with a defensible range, not a seller-defined point. The deal closed at £4.7m of attributed AI value — below the seller's claim, above the downside, and auditable line-by-line. Twelve months later, that same register is the core of the portfolio company's value-creation plan.
If you are running diligence on an AI-enabled target, the fastest way to see how this works is to run a small proof on a company you already know. Start with the Intangible Asset Valuator walkthrough to see the platform in action, then book a working session with the Opagio team to configure the AI-specific diligence branches for your deal.
You can also read the related pieces in this series — how PE firms use Opagio to maximise exit value, the Relief from Royalty method guide, and the intangible asset questionnaire guide — for the full picture of how structured intangible asset measurement changes deal economics.
AI is not going to stop being the most contested intangible category in M&A. The firms that get ahead of it are the ones that stop asking "is the AI real?" and start asking "what is it worth, under what method, with what evidence, and what range?" That is the conversation the platform is built to have.
Assess your 12 value drivers to understand your intangible asset strengths, gaps, and growth opportunities.
Private equity firms are acquiring AI-enabled targets at record rates, but traditional due diligence frameworks miss critical AI-specific risks. This article provides a comprehensive AI due diligence methodology covering technical, commercial, regulatory, and talent dimensions.
Read more →
AI washing inflates valuations and distorts capital allocation. This guide provides a structured due diligence methodology for detecting exaggerated or fabricated AI claims during investment evaluation, with specific technical and commercial red flags.
Read more →
Mergers and acquisitions involving artificial intelligence companies fail at a 62% rate in terms of value creation. The root cause is not market misjudgment — it is due diligence frameworks designed for physical assets, applied to intangible-asset-heavy technology companies. Here is the anatomy of that failure, and a framework that fixes it.
Read more →Get the latest insights on intangible asset growth and productivity delivered to your inbox.
Book a free consultation to see how Opagio Intangibles can help your business.