The AI Readiness Diagnostic Methodology

6 dimensions. 19 capabilities. 54 building blocks. 3 independent evidence streams. 118 prerequisite dependency rules.

The structure behind the five deliverables.

WHY METHODOLOGY MATTERS

The quality of an AI readiness diagnostic is determined by the quality of its methodology.

A diagnostic that evaluates six dimensions without sub-structure produces a directional score useful for general orientation, inadequate as a basis for investment decisions. A diagnostic that resolves those dimensions into capabilities and building blocks, that models the prerequisite relationships between them, and that triangulates evidence from multiple independent sources produces something different. It produces a sequenced plan.

The difference between these two outputs is commercially significant. A directional score tells the board that the organisation is "developing" on governance. A sequenced plan tells the board that governance is blocked by three specific data foundation gaps, that addressing those gaps also unlocks four other downstream capabilities, and that the right order of investment is data foundations first, governance second, use case pipeline third.

SIX DIMENSIONS

The six dimensions of AI readiness.

The methodology evaluates AI readiness across six dimensions. Each dimension measures a different organisational capability that determines whether AI investment will deliver value.

Strategy and Vision. Measures whether the organisation has a clear, leadership-owned view of what AI means for its business — beyond generic ambition. Covers AI vision and ambition, business case discipline, investment governance, competitive positioning, and leadership engagement with AI-specific decisions.

Data. Measures whether the organisation has the data foundations to support AI at scale. Covers data quality, data architecture, data governance, master data management, data integration, data accessibility, and the specific data readiness requirements of generative and agentic AI use cases.

Processes. Measures whether the organisation's core business processes are documented, standardised, and AI-ready. Covers process documentation, process standardisation across units, decision standardisation, process performance measurement, and the specific process requirements that determine whether AI can be meaningfully integrated.

Technology. Measures whether the technology foundations are in place to support AI deployment and scale. Covers core business systems, API and integration architecture, cloud and compute readiness, legacy system management, MLOps and LLMOps capability, and AI-specific infrastructure.

Skills and Capabilities. Measures whether the organisation has or can credibly build the talent, skills, and change capacity required for AI adoption. Covers technical AI capability, applied AI practitioners, organisational AI literacy, change management capability, AI-specific workforce planning, and leadership digital fluency.

Governance. Measures whether the organisation has the governance structures required to deploy AI responsibly and scale it safely. Covers AI policy framework, AI risk management, regulatory compliance (including the EU AI Act), ethics and bias controls, AI incident response, and board-level AI oversight.
 

CAPABILITIES AND BUILDING BLOCKS
 

Each dimension resolves into capabilities and building blocks.
 

The six dimensions are where most AI readiness assessments stop. The Perform Advisory methodology goes two levels deeper.
 

Each dimension resolves into three or four capabilities specific organisational competencies that a diagnostic can evaluate independently. Each capability then resolves into four or five building blocks — the concrete, assessable constituents of the capability.
 

In total, the methodology evaluates nineteen capabilities and fifty-four building blocks across the six dimensions.

Why this resolution matters. At dimension level, "the organisation scores three out of five on Data" is directionally useful and operationally hollow. At building block level, findings become specific and actionable — which master data practices are mature, which integration gaps are constraining AI use cases, which governance controls exist only on paper, which capabilities must strengthen before downstream investments pay off.
 

For the Data dimension as a representative example, the methodology assesses capabilities such as Data Foundations, Data Quality, Data Governance, and Data Enablement for AI — each containing specific, individually-assessed building blocks. The full map of capabilities and building blocks across all six dimensions is proprietary to the methodology and is walked through with clients during the Diagnostic.

Each building block is assessed against a five-level maturity scale, calibrated by role and triangulated across evidence streams. The organisation does not receive a single dimension score — it receives capability and building block scores that surface exactly where strength sits, where gaps sit, and where the next investment should go.

THREE-STREAM EVIDENCE TRIANGULATION

Three independent evidence streams. One triangulated picture.
Most AI readiness assessments rely on a single evidence stream — typically a self-assessment survey, sometimes supplemented by interviews. This is methodologically fragile for one reason: self-reported data is the least reliable kind of readiness evidence.

The Perform Advisory methodology draws on three independent streams and treats them as hierarchically ranked.

Stream 1 — Structured online assessment. Role-calibrated questions distributed to executives, managers, and operational staff. The assessment uses anchor questions to identify perception gaps between organisational layers — the gap between what executives believe and what operational staff experience is itself diagnostic evidence. Assessment data establishes the baseline score. It is the broadest but least reliable stream.
Stream 2 — Structured stakeholder interviews. Pre-mapped conversations with key decision-makers and domain owners, where each question traces directly to specific building blocks of the methodology. Interview evidence is not used only for narrative colour — it produces structured, scored evidence that validates, contradicts, or refines the assessment scores. Interviews are the second-ranked stream.
Stream 3 — Document review. Request and review of critical organisational documents, including AI or digital strategy, process documentation, data governance policy, and IT architecture. Documents are ranked as the highest-trust evidence stream in the methodology. The reasoning: what an organisation has actually documented, maintained, and kept current is a more reliable signal than what its people believe or describe. Stale documents are evidence of intent without execution. Absent documents are evidence of maturity gaps. Current, specific, operational documents are the strongest signal of capability.

What happens when streams conflict. The methodology has explicit rules for resolving evidence conflicts. Document evidence prevails over interview evidence. Multi-role interview consensus prevails over single-role claims. Operational-layer evidence is weighted against executive-layer claims for blocks where the executive view is known to be optimistic. Every scored block in the final output carries a confidence indicator based on evidence stream convergence.

This structured triangulation is what allows the Diagnostic to produce findings that withstand boardroom scrutiny. When the CFO asks "how confident are you in this number," there is a defensible answer.

PREREQUISITE DEPENDENCY MODELING

Readiness dimensions are not independent.

Most diagnostics treat readiness dimensions as separate columns on a scorecard. In practice, they are deeply interdependent.
Data quality gates AI governance — without reliable data, there is nothing meaningful to govern. Process standardisation gates automation — without consistent processes, automation amplifies chaos rather than reducing it. Skills gate scaling — without applied AI practitioners, pilots cannot progress to production. Governance gates regulatory-exposed use cases — without AI risk management, high-risk applications cannot be deployed.

These dependencies matter because they determine the order in which investments pay off.
The Perform Advisory methodology models one hundred and eighteen specific prerequisite relationships between building blocks across the six dimensions. These rules encode which building blocks must reach a given maturity level before specific downstream blocks can progress.

THE FIVE DELIVERABLES

Each deliverable is designed for a specific stakeholder and a specific decision.
 

The Diagnostic produces five deliverables. Each is shaped by who will read it and what they need to decide.
 

Diagnostic Report. The full analytical output. Sixty to eighty pages. Contains the complete scored picture across all 54 building blocks, the perception gap analysis, the dependency analysis, and the detailed findings. Designed for the executive sponsor, the CTO or CIO, and anyone who will own the subsequent work. This is the reference document.
 

Executive Summary. Ten to fifteen pages. Distils the report into what the CEO and board need to know to make investment decisions. Leads with the investment case — what current gaps cost, what Phase 1 investment is required, what addressable value becomes available. Designed to be read in one sitting.
 

C-Suite Presentation. A board-ready slide deck, typically forty to fifty slides. Structured to walk a leadership team through the findings in ninety minutes, with clear decision points throughout. Designed to be used by the sponsor to brief the full leadership team.
 

Sequenced Roadmap. The twelve-to-eighteen-month action plan that comes out of the dependency analysis. Organised by phase, with named owners, expected outcomes, and dependencies between actions made explicit. Designed to move the organisation from insight to action.
 

90-Day Quick Win Plan. The Monday-morning document. Three to five specific actions that can begin within the first week after engagement close, each with a named owner and a first-week deliverable. Designed to convert diagnostic momentum into organisational movement.

Each deliverable traces back to the same triangulated evidence base. What changes between them is the level of compression, the stakeholder, and the decision they enable.

WHAT HAPPENS AFTER THE DIAGNOSTIC

The Diagnostic is designed as the first engagement in a broader journey.

The methodology is rigorous enough to stand alone. Clients who need only a diagnostic — for board approval, investment committee preparation, PE portfolio review, or regulatory preparation — receive a complete deliverable set and conclude the engagement there.

Most clients, however, continue into follow-on work. The typical path:

• AI Strategy and Roadmap Development — converting findings into a prioritised, phased strategy with investment sequencing, use case selection, and governance design

• POC Design and Delivery Oversight — scoping, designing, and overseeing proof-of-concept initiatives that convert strategy into live evidence of AI value

• AI Operating Model and Governance Design — designing the target operating model, decision rights, committees, and policies required for AI to scale safely

• Change Management and Adoption Support — leading the organisational work that determines whether AI investments are actually used and sustained
   

Technical implementation is delivered through our vetted network of AI engineering and IT partners. Perform Advisory remains in the strategy and oversight lane throughout — this keeps our advisory independent, our findings unconflicted, and our clients free to choose the right implementation partner for their specific context.

HOW TO TELL VALUABLE DIAGNOSTIC FROM A TEMPLATE

Six questions to ask any AI readiness provider.

We publish this section because we think the market would be better if buyers asked these questions more often.

1. How many building blocks does your methodology actually evaluate? If the answer is the number of dimensions (six, seven, eight), the methodology operates at dimension level only. This produces directional scores, not sequenced plans.

2. Do you model dependencies between building blocks? If the answer is no, the roadmap you receive will be consultant judgment about what to do first, not a systematic analysis of which investments unlock which downstream capabilities.

3. How many independent evidence streams do you triangulate? If the answer is one (self-assessment survey), the findings rest on self-reported data. If the answer is two (survey plus interviews), document reality is not being checked. Three streams is the minimum for defensible diagnostic work.

4. What happens when your evidence streams conflict? If there is no structured rule for resolving conflicts, the consultant resolves them by judgment. This is fine — but ask how the consultant decides, and whether their decisions are consistent across clients.

5. Who delivers the engagement? In firms using a partner-sells, analyst-delivers model, the person who presented in the pitch is rarely the person doing the work. Ask who specifically will be on your project, and at what seniority.

6. Do you sell technical implementation? Firms that do have a structural incentive to surface problems their own teams can solve. This does not mean they will — but it is worth understanding the commercial architecture of the advisor before hiring them.
These are the questions we wish every client asked before they signed a contract — including, in many cases, with us.