AI readiness
It takes whatever your organisation already does and does more of it, faster. Most companies asking which AI tools to buy should ask these six questions first.
Stage 1 · Decide
The team knows what matters and can say no to everything else.
The three-answer test. Ask three people, separately, to name the top three priorities this quarter. Reveals:Three different answers mean the team does not share priorities.
AI speeds up delivery. If the priorities are wrong, you just build the wrong things faster.
Stage 1 · Decide
One named owner per outcome, with decision-making authority.
The failed-project test. Pick a recent project that went wrong and ask who owned it. Reveals:A team name means ownership is diffuse. Multiple names mean it is contested.
AI makes producing work easy. Someone still has to review it. If no one owns it, it piles up unread.
Stage 2 · Learn
Retrospectives happen, produce actions, and the actions change behaviour.
The action audit. Pull the actions from the last three retros and count how many were done. Reveals:If most actions are still open, the retros are not changing how the team works.
AI makes mistakes. Someone has to check its work and fix the causes. If no one does, the same mistakes keep happening.
Stage 2 · Learn
How the organisation works is written down, not held in one person’s head.
The bus-factor count. For each critical system, ask who else could run it tomorrow. Reveals:Every answer that comes back ‘nobody’ is a dependency AI cannot help with.
AI is only as good as the context you give it. If how things work lives in people’s heads, there is nothing to point it at.
Stage 3 · Scale
Trusted data, trusted deployments, and clear rules on what AI is allowed to do.
The metric trace. Take one number leadership reviews and trace it back to raw data. Count the manual steps and name who would notice if one broke. Reveals:Fragile manual pipelines that any AI application will inherit.
AI gives confident answers whatever the data behind them. If the data is bad, the answers are wrong at scale and hard to spot.
Stage 3 · Scale
The organisation has adopted new ways of working before, and made them stick.
The adoption ledger. List the last five changes you attempted and count how many stuck past six months. Reveals:Fewer than three, and AI is just another change that will not stick.
Adopting AI changes how people work. Push it on a team that is not ready and people will quietly stop using it.
Score each 1 to 4
At that level AI makes these gaps worse. A small pilot is a good way to find them. Do not move core operations onto AI until you have closed them.
Our diagnostic covers these six dimensions. We spend time in the business, run the field tests and score each one, then give you a plan for the next few months or quarters. If any of these were hard to score, get in touch.
We have built our own tools for two of the six: feedback loops and documentation. Huw explains why in Everyone is building company brains.