How to Run an AI Readiness Assessment on Your Business

Short answer: Assess five dimensions — data quality, process definition, technical foundations, team capability and governance — and score each honestly. A business weak on data and process will not get value from AI regardless of how good the tools are. Readiness is mostly about foundations, not about AI knowledge.

Dimension 1: Data

The dimension that most determines outcomes.

• Is there one authoritative record per client, project and transaction, or do duplicates exist across systems?
• Are the fields you'd want AI to analyse structured, or free text?
• Can you export your operational data in a usable format?
• Do you know what personal data you hold and where?
• If you sampled twenty records right now, how many would be accurate?

Weak here means: stop. Fix the data layer before deploying anything. AI on dirty data produces confident nonsense — fluently, at speed, with no error message.

Dimension 2: Process

• Are your core processes documented, or do they live in people's heads?
• Do people execute the same process the same way?
• Can you identify which steps are rule-based and which need judgement?
• Do you know how long each process actually takes?

Weak here means: you can't identify good use cases, because you don't have a clear picture of where time goes. Map first.

Dimension 3: Technical foundations

• Do your core systems have APIs, and do you know whether they're usable?
• Is there an automation platform in place, or would AI need to be operated manually?
• Who maintains integrations when they break?
• Is authentication and access managed centrally?

Weak here means: AI stays a manual tool — someone copying text into a chatbot — rather than becoming part of a workflow. That's still useful, but the gains are a fraction of the potential.

Dimension 4: Team capability

• Has anyone had structured training on AI tools, or is it self-taught and uneven?
• Does the team understand what these tools get wrong, not just what they do well?
• Is there someone who can own an AI-assisted workflow?
• Is there appetite, or resistance? Both are legitimate and both need handling.

Weak here means: deployment will stall at adoption. This is also where the AI literacy obligation bites — training isn't only good practice, it's a legal requirement for organisations in scope.

Dimension 5: Governance

• Is there an AI usage policy naming approved tools and data rules?
• Do you have a register of AI systems in use, including tools adopted without approval?
• Have you assessed whether any use case falls into a regulated category?
• Are DPAs in place with AI vendors processing personal data?
• Is there a review step defined for AI output?

Weak here means: you're carrying unquantified risk. This is fixable in weeks, unlike the data dimension.

Scoring and what to do with it

Score each dimension one to five. Then:

• Data or process below three — foundations first. AI deployment now will produce disappointing results and you'll wrongly conclude AI doesn't work for your business.
• Foundations solid, technical or governance weak — start with contained, human-reviewed use cases while you close those gaps.
• All dimensions three or above — pick one high-value use case, deploy properly with measurement, then extend.

Frequently asked questions

How long does an assessment take?

A day or two internally. The uncomfortable part is honesty about the data dimension, which most businesses overestimate.

What's the most common weak spot?

Data, consistently. Second is governance — plenty of AI use, no policy, no register.

Can we deploy AI while fixing foundations?

Yes for standalone uses like drafting and summarising, which don't depend on your data. No for anything reading your business records.

Our systems audit covers all five dimensions. Book a free assessment.