The board asks you for an AI strategy. On the ground, you have a handful of ChatGPT licences and a few enthusiasts tinkering on their own. In between, nobody has a clear view of what AI could actually change in your company.
That is what an AI audit is for. Knowing where AI creates value for you, with your processes and your teams. Not in some case study you read on LinkedIn.
Here is the method: where to start, what to map, how to prioritise, and how to keep your audit from ending as a report in a drawer.
Start by scoping, not by scanning everything
The first reflex to avoid is sending a questionnaire to the whole company. You will collect hundreds of vague answers and nothing you can use. Start from a business question instead: where are you losing time or margin today? That is where your audit begins.
Then find a sponsor. Someone in leadership who will carry the decisions the audit produces. Without a sponsor, your report ends up as an attachment nobody opens. Pick the functions to look at first too: the ones where teams are overwhelmed and repetitive tasks keep piling up.
Good scoping fits on a single page: the objective, the scope, the sponsor, the people to interview. If you cannot write it down, you are not ready to audit.
Map the processes, not the tools
The number one mistake in AI audits is listing the tools. Copilot or ChatGPT, which model, which licence. Wrong way in. A tool sitting on a fuzzy process is just one more subscription on the bill.
What you map is tasks. Go to each function and ask: what eats your time every week? The answers are always specific. Writing up meeting notes, responding to tenders, sorting through applications, producing the monthly reporting.
For every task you spot, note four things: how often it happens, how long it takes, who does it, and whether the data it needs is actually accessible. That grid is enough to move on to prioritisation.
Prioritise with the impact x feasibility matrix
Now you have your list of tasks and processes. Two questions per opportunity. Impact: how much time or value do you free up if AI takes over? Feasibility: is the data accessible, and is the team up for it?
Drop each opportunity into one of four boxes. High impact and high feasibility: your quick wins, launch them now. High impact but hard to pull off: your structural projects, plan them. Low impact and easy: let the teams handle it. Low impact and hard: forget it.
That matrix turns an inventory into a plan of attack. It is what separates a snapshot from an action plan.
The four traps that sink an AI audit
First trap: trying to audit everything. Going for exhaustive gets you a doorstop nobody reads and teams worn out before they have launched a single thing. A few functions dug into properly beats the whole company skimmed.
Second trap: staying at tool level. Comparing licences tells you nothing about how you create value. Third trap: auditing without the teams. The people who do the work know where the hours go, the org chart does not.
Fourth trap: the audit with no follow-through. If the deliverable commits no one, with no owners and no deadlines, you paid for a photo. Nice to look at, useless.
The final deliverable: a roadmap you can act on
A good AI audit ends with four deliverables: a map of the high-impact AI opportunities in your organisation, a set of quick wins, a clear prioritisation, and a function-by-function roadmap. Every team knows what to do, in what order, toward what goal.
That is exactly what the AI Studio at AI x Leaders produces: a structured diagnostic of 4 to 8 weeks that maps the high-impact opportunities, identifies the quick wins, prioritises them, and delivers an actionable roadmap function by function.
Not a theoretical doorstop. A concrete action plan, with priorities you actually own and quick wins named function by function.
Frequently asked questions
How long does an AI audit take?
It depends on the size of the organisation and the scope you choose. At AI x Leaders, the AI Studio diagnostic runs over 4 to 8 weeks, mapping and roadmap included.
Who should you involve in an AI audit?
A sponsor in leadership, and the people who do the work day to day. They are the ones who know the repetitive tasks and the lost hours, not the job descriptions.
Can you run an AI audit in-house?
Yes, if you have the time and someone able to challenge every function without going soft. An outside eye mainly brings prioritisation free of internal politics, and a comparison with what works elsewhere.
What is the deliverable of an AI audit?
A map of high-impact opportunities, a set of quick wins, a clear prioritisation, and a function-by-function roadmap. If all you get is a state of play, the work is not finished.
Want to know where AI creates value in your organisation? Our AI Studio runs the diagnostic in 4 to 8 weeks: a map of opportunities, quick wins, prioritisation, and a function-by-function roadmap.
Explore the AI audit →
