"Are we ready for AI?" is the question every board is asking, and most answers are either a vendor's pitch or an internal optimism survey. Readiness is not one thing. It is the state of your data, the systems that hold it, the skills of your people, the way your processes run, and whether anyone is accountable for the result. This checklist covers all five.
011. Data
- Do you know where your important data lives, and who owns each source?
- Is it accurate enough to act on? Pick one master (customers, materials, vendors) and sample it honestly.
- Is it accessible — can a system read it without a six-week integration project?
- Is it allowed to be used? Contracts, consent and regulation decide what any AI may touch.
Red flag: nobody can say how many duplicate records a core master holds.
022. Systems
- Which systems are the source of truth, and which are copies?
- Can they be integrated with — are there interfaces, exports, or only screens?
- Where would an AI system run: on your infrastructure, in a private cloud, or nowhere yet?
- What is the state of security and access control today? AI inherits whatever you have.
Red flag: critical processes depend on spreadsheets that live on one person's laptop.
033. People
- Who in the organisation can judge whether an AI's output is right? Domain expertise is the scarce resource, not technology.
- Who will run and maintain a system after the vendor leaves?
- Has anyone been given time to learn, or only tools?
- Is there a sponsor in leadership who will own the outcome?
Red flag: enthusiasm at the top, no capacity in the middle, no time at the bottom.
044. Process
- Which processes are repeatable enough to be improved, and which are improvised every time?
- Where does decision-making actually happen — in the system, or in a conversation after it?
- What would an AI's recommendation have to look like for someone to act on it?
- Is there a way to measure the process today, so improvement can be shown?
Red flag: the process exists in a document nobody follows.
055. Governance
- Who decides what AI may be used for, and what it may not?
- How will decisions made or influenced by a system be explained and audited?
- What happens when it is wrong — who notices, who corrects, who is accountable?
- Is there a policy on data leaving the organisation through AI tools, including the ones employees already use?
Red flag: staff are already pasting company data into public assistants and nobody has written down whether that is allowed.
06Scoring it honestly
Score each area from one to five, with evidence for each score, not impressions. Readiness is the lowest score, not the average: a company with excellent data and no one to run a system is not ready. The purpose is not a number to present, it is to know which area to fix first.
07What to do with the result
- Fix the lowest area before buying anything
- Choose a first project that your readiness can actually support, in a process you can measure
- Write the governance down before the first system goes live, not after the first incident
- Reassess in a year; readiness changes
In short
- Readiness is five things — data, systems, people, process, governance — and is measured by the weakest.
- Domain experts who can judge an output are the scarce resource, not technology.
- Score with evidence, fix the lowest area first, and write governance before go-live.
Questions
Can we do this assessment ourselves?
You can run the checklist internally, and it is worth doing. An outside assessment adds what internal ones rarely have: an honest answer with no stake in the result.
Do we need to be 'ready' before starting anything?
No. You need a first project that your current readiness can support. Starting too big is the common failure; starting small in a measurable process is the common success.
What does an audit deliver?
A written readiness assessment, a threat assessment, an opportunity map and a roadmap, presented to leadership and handed over to act on with any partner.
Sounds like your problem?
Tell us about it. We'll say honestly whether the swarm can help, and what it would take.
More guides