One of the most asked-about things on our website is the AI strategy audit. The pitch is simple: a short, fixed-fee engagement that ends with a clear plan of where AI would actually pay off in your business. The reality is more interesting than the pitch, and worth a proper look.
This is a detailed walkthrough of what a two-week audit actually involves, end to end.
What the audit is for
The audit exists for one reason. Most owners we talk to know they should be doing something with AI but cannot tell, honestly, where it would pay off in their specific business. The internet is full of generic answers. The strategy decks are full of generalities. What they need is somebody who will sit down with their actual business and tell them, with real numbers, where to start.
That is the job of the audit.
What it is not
It is not a sales call. The audit is paid work. We do it whether or not you go on to build with us. About half the firms that do an audit end up building with us afterwards. The other half take the report and either build it themselves, hire someone else, or simply use the analysis to make sharper decisions on their own.
It is also not a generic AI strategy framework. We are not in the business of repackaging Gartner slides. The deliverable is specific to your business, with named workflows, real cost estimates, and a sequenced roadmap.
The two-week shape
The work runs over two calendar weeks, with a few discrete phases.
Days 1 to 3: Setup and document review
A short kickoff call with the owner or senior decision-maker. We agree the scope of the audit, the people we need to speak to, and the documents we will be looking at. Then we read.
What we read: any process documentation you have, recent management accounts at a high level (only what you are comfortable sharing), customer enquiry samples, the last few weeks of internal team comms if you are willing to share, and anything else that gives us a real picture of how the business runs.
If you do not have much in writing, that is fine. We capture it through conversations instead.
Days 4 to 7: Conversations
Structured interviews with three to five people across the business. Usually the owner, a senior operator, someone customer-facing, and someone in finance or admin. Forty-five minutes each, focused on three questions: what work is eating your time, where does it slow the business down, and what would change if it disappeared.
The conversations are confidential within the audit. We do not quote individuals to other team members.
Days 8 to 10: Analysis
We pull everything together. Map the workflows. Identify the time sinks. Identify the integration points and the data the agents would need. Cost-estimate each candidate project. Sense-check feasibility against current AI capability.
This is the unglamorous middle of the work, and the part where most generic strategy reports cut corners. We try not to.
Days 11 to 13: Drafting
Write the report. Then re-read it the next morning and rewrite anything that sounds like consultancy waffle. Then have somebody else on our team challenge the strongest recommendations.
Day 14: Delivery
A two-hour delivery session with the owner and any senior people they want present. Walk through the report. Take questions. Disagree productively where there is something to disagree about. Leave the room with a written deliverable everyone has read.
What the report actually contains
The deliverable is a written report, typically twenty to thirty pages, structured the same way every time.
- Executive summary. One page. The three or four most important findings, the recommended sequence, the rough total investment.
- Where time goes today. A clear picture of how the business currently spends its hours, by function and by activity.
- Top opportunities. Three to five workflows where AI would meaningfully help, ranked by impact-to-effort ratio. Each one has a clear problem statement, a proposed agent design, an integration map, a cost estimate, and an expected payback period.
- What we would not do. The places where AI is the wrong answer for your business right now, and what would actually help instead.
- A practical sequence. Quarter by quarter, what to build first, second, and later. With reasons.
- Compliance and data considerations. Specific to your sector and your data.
The report is yours. You can take it to your team, your board, your accountant, or another agency. We do not lock you into anything.
The point of the audit is not to convince you to spend money with us. It is to make sure that whatever you spend, with whoever you spend it with, has a real chance of paying back.
What it costs
The audit is fixed fee. The exact number depends on the size of the business and the depth of the work, but for most UK SMEs it sits in the low four figures. Always quoted up front, always honoured.
If you go on to build with us afterwards, we credit a portion of the audit cost against the first build. If you do not, no awkwardness.
What we ask of you
For the audit to work, we need a few things from your side.
- One named decision-maker, available for the kickoff and the delivery session.
- Access to three to five people for the interviews.
- A reasonable willingness to share what you actually do, including the bits that are not pretty.
- Honest answers, especially when we ask about things that are not working.
Two weeks of your senior people's attention, in fragments. Nothing more.
If this sounds useful
If you are at the point where you suspect AI could help your business but cannot tell where to start, this is what the audit is for. Tell us about your business and we will set up a short call to confirm whether the audit makes sense for you. If we do not think it would, we will say so.
For more on what the first 90 days of an AI rollout typically look like after the audit, our piece on a practical 90-day AI roadmap for SMEs picks up where this one leaves off.