On this page
How does an AI audit work?
- Agree what is in view. The business and the auditor name the teams and processes to be reviewed.
- Map how work runs. The auditor follows one enquiry, order or client from first contact to payment, step by step.
- List the tools. Every system is written down with the data it holds and what it connects to.
- Find where work gets stuck. Examples are details copied by hand, jobs that wait on one person, and enquiries with no owner.
- Price each problem. The auditor counts how often it happens and what it costs in a year.
- Rank the fixes. Some need AI. Some need plain automation. Some need a change of process and no software.
- Write the report. The business gets the findings, the order of work and a price range for each fix.
The phrase has a second meaning. It can also describe a check of an AI system itself, for accuracy, bias and risk. This page covers the first meaning: a review of a business before it buys or builds.
What can an AI audit tell you, and what can it not?
| Question | Can an audit answer it? | Why |
|---|---|---|
| Where do our time and money go? | Yes | It uses counts taken from your own business |
| Which problems suit AI, and which do not? | Yes | Each problem is matched to the simplest fix that works |
| What should we fix first? | Yes | Fixes are ranked by cost against gain |
| Where must a person stay in charge? | Yes | Money, legal commitments, complaints and regulated advice are marked for human review |
| Will the build produce a set result? | No | A report estimates. It does not guarantee |
| Are we meeting our legal duties? | No | That needs a qualified legal adviser |
An audit also fixes nothing by itself. It is a report. Its accuracy depends on the business sharing real numbers and access to its tools.
Terms you will meet
| Term | Plain meaning |
|---|---|
| Process map | A drawing of each step in a piece of work, and who or what does it |
| Bottleneck | A step where work waits and a queue builds |
| Integration | A connection that passes data between two systems |
| Scope | The written document that says what will be built, for what price |
| Human review | A point where a person must check or approve before the system acts |
Who needs one, and who does not
An audit fits a business in one of these positions:
- You are about to spend on AI or software and do not know where to start.
- You run several tools that do not share data.
- The team is full, and the next step looks like another hire.
- You can feel money leaking but cannot say which gap is the largest.
It is the wrong purchase in these cases:
- You have one small, obvious problem. If an off-the-shelf tool fixes it, buy the tool.
- Your business is new and the process changes each week. There is nothing stable to map yet.
- You only want a rough first view. Our free Business Systems Scorecard asks 16 questions and costs nothing.
What the Agentryx audit covers
Our audit is the first stage of every build. We map how your business runs, the tools it uses and where work gets stuck. You receive a written report within 10 working days.
The report names the fix for each problem. That may be workflow automation where staff copy details between tools, CRM setup and migration where enquiries have no owner, or dashboards and reporting where numbers are gathered by hand. It may also say that no build is needed.
The stages after the audit are design, build, test, launch and improve. Our approach page describes each one. For the kind of starting point an audit maps, see TMMB Academy, which came to us with fragmented processes, rising software subscriptions and no single system.
What does an AI audit cost?
Our audit costs £3,000 plus VAT in the UK and $5,000 in the US. It is paid in full.
The fee is credited against the build when the build deposit is paid within 14 days of the report. Scoped builds run from £5,000 to £20,000 (US $10,000 to $30,000). Before the audit we give a range, and the exact price follows the written scope.
Our AI audit pricing page sets out what is included.
What is it worth? Do the sum
The leak, in an owner's words: "We are losing money somewhere, and we do not know which gap is the biggest." An audit does not pull a lever by itself. It shows which of four levers to pull first: more clients, more money per client, money kept, or time back.
The mechanism is the pricing step. Each problem gets a count, a value and a yearly figure, laid out like this:
| Example problem | The sum | Cost a year |
|---|---|---|
| Enquiries with no next step | 8 a month × 1 in 4 close rate × £1,000 client value × 12 | £24,000 |
| No-shows | 2 a week × £75 appointment value × 52 | £7,800 |
| Copying details between tools | 5 hours a week × £20 an hour × 52 | £5,200 |
| Total | £37,000 |
Example figures. Swap in your own. They show the method. They are not a result, an average or what an audit will find.
To find your own numbers, count three things for one normal week. Enquiries that got no reply or no next step. Appointments nobody turned up to. Hours spent moving details from one tool to another. Then work out your own figure with our ROI calculator.
Bring those counts to a call with us. The audit report replaces every example above with a figure from your own business.
Frequently asked questions
How much does an AI audit cost?
Prices vary by provider and depth. Ours is a fixed fee, stated in the cost section above, and it can be credited against a build.
What do you get at the end of an AI audit?
A written report. It lists each problem found, what it costs in a year, the fix we recommend and the order to do the work in.
How long does an AI audit take?
Our written report is delivered inside the 10 working day window described above. The build timeline is separate and is agreed in the written scope.
Do we have to use AI after an AI audit?
No. Some problems are better fixed with plain automation or a change of process. A good audit says so.
Is an AI audit the same as a free consultation?
No. A consultation is a conversation. An audit is paid work that maps your processes and tools and ends in a written report.
What do you need from us for the audit?
Time with the people who do the work, access to the tools they use, and real numbers such as enquiry counts and hours spent.