Your Data Is Better Than You Think, and Worse Than You Think

Ask a leadership team how good their data is and you'll usually get a single number: a 3 out of 5, an "okay, not great", a nod towards the data warehouse. That number is the wrong unit of measurement. Data quality is a different fact for every domain, every system, and every person who has to extract from it, and the AI Readiness Assessment (03a) inside Section 3: AI Readiness Assessment only earns its keep when you resist collapsing it into one figure.

The Data That's Better Than You Assumed

A grocery operator running 80-plus stores came into the assessment expecting a low result across the board. What they found was a split running the other way. Their technical scores were strong: data quality 3.5, digital maturity 3, governance 3. Their organisational scores were weak: culture 2, leadership alignment 2, skills 1.5.

The instinct in a business that feels behind on AI is to assume everything is behind. This operator's commercial data foundation was already in reasonable shape, and the constraint sat somewhere else entirely: whether the organisation had the appetite and alignment to do anything useful with what the data showed. That's a different problem, with a different fix, and confusing the two would have sent the programme after a data remediation project it didn't need while leaving the actual blocker untouched.

The Algorithm Maturity Level (03d) tool is useful here precisely because it separates the data-and-technology question from the people-and-leadership question. A business can sit comfortably in Descriptive or Diagnostic analytical maturity, which is a perfectly workable starting point for AI experimentation, while still needing months of leadership alignment work before anyone acts on what the data shows.

The Data That's Worse Than You Assumed

A PE-backed DTC brand ran the assessment under pressure. The sponsor wanted it completed in a week; the programme team pushed back and took three. What came out the other side was a leadership team genuinely aligned (a score of 4, sponsor and founder on the same page) sitting alongside a data quality score of 1.5 and a digital maturity score of 2.

Confidence at the top doesn't repair a data estate. A small, fast-moving team with a founder and a PE sponsor who agree on strategy can still be running on data that isn't there yet, and the risk in that situation is specific: the same alignment that makes a business feel ready is what makes a low data score easy to wave past. Nobody in the room is arguing against the AI plan. Nobody is checking whether the data underneath it can carry the weight.

That's what a Readiness Guide: Data Foundations (03e) workstream is for: a parallel track that closes the gap between what leadership believes and what the data can support, rather than a reason to put the whole programme on hold until the data catches up.

Ask "Data For What," Not "How's Our Data"

The pattern across both businesses is the same. A single data quality score answers a question nobody is asking. The narrower question is the one that matters: how good is the data behind the specific decisions this AI programme needs to make?

Scoring data quality against the domains where the planned AI use cases sit turns that number into a planning input a steering group can act on.

The AI Readiness Facilitation Guide (03b) exists to force that specificity. When scores diverge by more than a point between assessors, the divergence is itself the finding, and it deserves investigation before anyone averages it away. The follow-up question is always the same one: which domain does this AI programme depend on, and whose score describes that domain? Once you have separate, evidenced answers for the domains that matter, the AI Readiness Profile (03c) turns them into a shape a steering group can act on.

A business that knows exactly which domain is strong and which is weak has more useful information than one that knows its "data" scored a 3. The second number is comfortable. The first one is usable.