What is AI readiness?

The organisational ability to deliver, operate and govern an AI-supported workflow: data access, process stability, integration, skills, governance and change capacity.

Can a company with poor data readiness still use AI?

Yes, by choosing workflows whose inputs are already digital and self-contained — inbound email, documents, ticket text — while broader data work happens in parallel.

AI Strategy

AI readiness assessment: what to measure, and what to do about the gaps

The dimensions of AI readiness — data, process, systems, skills, governance and change capacity — how to score them, and how readiness should influence sequencing.

Miguel Torres, Founder, Merjora · Published 24 August 2026 · Updated 3 September 2026

In short

An AI readiness assessment scores six dimensions: data availability and quality, process stability, systems and integration access, skills and capacity, governance and risk control, and change capacity. Readiness does not determine whether an opportunity is valuable — it determines how quickly it can be delivered and how much of the first project is preparation. Low readiness on one dimension usually delays a project; low readiness on data plus governance together usually means starting somewhere else.

Key takeaways

Readiness constrains sequencing; it does not veto value.
Process stability matters more than data volume for most first projects.
The scarcest readiness resource is reviewer time, not engineering time.
Score readiness per opportunity, not once for the whole company.

The six dimensions

Dimension	Strong signal	Weak signal
Data	Inputs exist digitally in a reachable system	Data lives in email attachments and personal drives
Process	Steps are documented and consistent across the team	Every person does it slightly differently
Systems	APIs or exports exist and someone owns them	Integration requires a vendor change request
Skills	Someone can evaluate model output critically	Nobody can say whether an output is good
Governance	Data protection and oversight positions exist	No classification, no named owner
Change capacity	The team has bandwidth to adopt and review	Team is already at capacity with no reviewer time

What to do with a low score

Low data readiness: start with a workflow whose inputs are already digital, and fix data access in parallel.
Low process readiness: document and standardise the process first; automation of an inconsistent process amplifies the inconsistency.
Low skills readiness: build the evaluation capability during the first project rather than before it.
Low governance readiness: classify one use case properly and reuse the pattern.
Low change capacity: reduce scope until reviewer time is realistic, or delay.

Readiness and value together

Merjora scores readiness alongside opportunity value so the roadmap reflects both. A high-value opportunity with a readiness constraint is not removed from the portfolio — it is sequenced after the constraint is addressed, with the constraint named explicitly.

Frequently asked questions

What is AI readiness?: The organisational ability to deliver, operate and govern an AI-supported workflow: data access, process stability, integration, skills, governance and change capacity.
Can a company with poor data readiness still use AI?: Yes, by choosing workflows whose inputs are already digital and self-contained — inbound email, documents, ticket text — while broader data work happens in parallel.

Score your readiness alongside your opportunities

Merjora's assessment returns a readiness profile with the specific constraints that affect sequencing.

Assess your AI readiness