Most AI-readiness audits are theater. A consultant turns up, asks twenty questions, runs them through a maturity matrix, and hands you a forty-page deck rating you yellow on six dimensions you have never heard of. You pay a lot, you read a little, you build nothing different. We have read those decks too. We do not write them.
Our audit is three working days. At the end you get a written report, somewhere between eight and twelve pages, that says: here is what is actually happening at your company today, here is the one thing you should ship first, here is the one thing you should not ship at all, and here is what each path costs in time and money. The CEO reads the whole thing in a single sitting. That is the point.
The three-day shape
Day one is interviews. We talk to the owner or CEO, we talk to whoever runs operations, and we talk to one or two front-line people who actually answer customer questions. Three to five conversations, forty-five minutes each, recorded with permission. We are not running a survey. We are listening for the gap between what leadership thinks the business does and what the business actually does on a Tuesday afternoon.
Day two is the technical pass. We look at your website the way an AI assistant would look at it. We pull your sitemap, your robots.txt, your structured data, and your existing knowledge assets if you have any. We open ChatGPT, Claude, Perplexity, and Gemini, and we ask each of them seven to ten questions a real customer would ask about your business. We screenshot what comes back. We record what was right, what was wrong, and what was simply absent.
Day three is synthesis and writing. We map every finding to one of five domains (more on those in a moment), score each domain on a one-to-five scale we will explain in the report, and write the recommendations. By end of day three you have the draft. We sit with you for a forty-five-minute readout, take notes on anything you push back on, and ship the final version within two business days.
That is the whole shape. No deliverable theater, no twelve-week consulting engagement, no three-figure hourly rates billed against a retainer. The audit is fixed-price, fixed-scope, fixed-time.
The five domains we look at
We do not measure abstract AI maturity. We measure five concrete things, every time, in every audit. Together they answer the question: if a customer or a procurement team or a journalist tried to find out about your company today, with the tools they actually use, what would they find?
1. Visibility to AI assistants
We count the customer-facing surfaces of your business that an AI assistant can actually read. Most SMBs have between three and five of these (homepage, about, services, pricing if they have it, contact). Most of those surfaces are written for humans scanning at speed, not for an assistant trying to extract a clean answer. We measure two things: can an assistant find your content, and can it parse what it found into a useful answer. Almost always, the gap is on the second part.
2. Tooling and stack hygiene
What does your team actually use day to day? We ask for the list, we look at how the tools talk to each other (or do not), and we note the seams where work falls through. The point is not to recommend a stack rebuild. The point is to know what is real before we recommend anything new.
3. Data hygiene
How is your customer data, your product data, your operational data stored? Is it queryable? Is it exportable? Is it duplicated across four spreadsheets and a Notion page? Most SMBs have a single source of truth on paper and four sources of truth in practice. We map them.
4. Team and ownership
Who owns AI-related decisions at your company today? Most of the time, the honest answer is: nobody. The CEO is interested but busy, the operations lead is overloaded, and the marketing team has tried two or three tools and given up. We note this, because recommendations without an owner are recommendations that do not ship.
5. Customer-touch surfaces
Where does your business actually meet a customer? Email, phone, chat, in-person, a form on the website? For each surface, we ask: can a recommendation we make actually reach the customer there, or does it die on the way? This is the domain that most audits skip, and it is the one that most often determines whether a project ships or stalls.
What you get at the end
One PDF. Eight to twelve pages. The structure is consistent across every audit we run, which makes it easy to compare year over year if you re-audit later.
The first page is the executive summary, written for the decision-maker who has fifteen minutes. It says, in plain language: this is your current shape, this is the first thing you should build, this is the second thing, and these are the things you should not build (and why). Most readers stop here. That is fine.
Pages two through six are the five-domain breakdown. For each domain, we score one to five, we list the three or four findings that drove the score, and we write a paragraph on what good would look like. No charts that say nothing. No fluff.
Pages seven through ten are the recommendations. Three or four specific, scoped projects, in priority order, with rough cost and time estimates, what we think the business case is, and what could go wrong. We include at least one project we do not recommend you do (and why), because half the value of the audit is permission to not do the thing you were already worried about.
The last page is appendix: the questions we asked the AI assistants, the answers we got back, and the screenshots. This is the receipts page. If you ever want to ask us “how do you know our visibility is bad,” that page is the answer.
What it costs vs. what skipping it costs
The audit is the cheapest thing we sell. It is also the one that changes the most. We have run audits where the recommendation was: do not start the chatbot project you were planning, fix three pages on your website first. The chatbot project would have cost forty thousand and shipped a confused bot. The page fixes cost a fraction of that and moved more revenue. We are not in the business of upselling. We are in the business of telling you the truth about where you are.
What does skipping the audit cost? Almost always it is one of two things. Either you spend money on the wrong project (the classic: “let’s add an AI chatbot to our site” before any of the content the bot would draw from is in good shape), or you spend no money at all and a competitor moves first because they had a clearer picture of what to do. Both of those costs dwarf the cost of the audit. That is not a sales pitch. It is what we have watched happen, repeatedly, across every business we have worked with.
What we expect from you
Three things. First, three to five hours of time across the right people, spread over the three audit days. Second, read access to whatever data we ask for (your CRM exports, your analytics, your existing content). Third, honesty during the interviews. The audit is only as good as the inputs, and the inputs are the conversations.
We do not ask for a slide deck about your strategy. We do not ask for a NDA before the first conversation (we sign one before we touch the data). We do not ask you to fill out a fifty-question intake form. We have done that work; you have not, and your time is the constraint.
Ready to find out where you actually are?
The audit is the first thing we ship for almost every new engagement. If you are not sure whether you need it, you probably do. Read the visibility service page for what comes after, or get in touch and tell us what you have tried so far. We will tell you whether an audit is the right next step, or whether you should skip it and build the thing you already know you need.