What actually happens in an AI readiness audit
"Readiness audit" sounds like the kind of thing a consultant invents so they can charge you before doing any real work. Fair suspicion. So let me tell you exactly what we do during one, why each part matters, and what you walk away with.
The short version: we spend a week or two figuring out whether AI will actually help your business right now, where it would help most, and what needs to be true before it works. Sometimes the most valuable thing we deliver is "not yet, and here's why."
Why bother auditing first
Plenty of businesses skip this and jump straight to buying a tool. A chunk of them end up with software nobody uses, or an automation that quietly produces wrong answers for three months before anyone notices.
The audit exists to prevent that. AI is unusual in that it fails confidently. A traditional bug crashes and you know something's wrong. A badly set up AI assistant keeps answering, just incorrectly, in complete sentences. You won't catch it from a status dashboard. You catch it by understanding, up front, where your information is reliable and where it isn't.
What we look at
There are four things we dig into. None of them are glamorous.
Your data
This is where most of the audit's time goes. We trace where your important information lives and what shape it's in. Customer records, product details, pricing, past orders, support history. Is it in one system or seven? Is it current? When two systems disagree, which one is right, and does anyone actually know?
We're not judging. Every business we've worked with has messy data somewhere. We just need an honest map, because that map determines what's possible. A company with clean order history can do useful things with it next month. A company whose order history is split between a legacy database and a filing cabinet has a different first step.
Your processes
Next we follow a few of your real workflows end to end. Not the version in the procedure manual, the version that actually happens, including the workarounds and the "oh, for that one we just email Dave."
This matters because AI automates steps, and you can't automate a step nobody can describe. Often we find that a process people called "automated already" has three manual handoffs buried in it. Sometimes we find a process so tangled that the right recommendation is to simplify it before adding any technology at all.
Your team
Tools don't adopt themselves. We talk to the people who'd actually use whatever gets built. Are they drowning and desperate for help, or are they wary that this is a prelude to layoffs? Who's curious? Who's going to quietly refuse to change how they work?
This isn't a soft, optional part. The best-built automation in the world dies if the team routes around it. Understanding who your allies and skeptics are shapes how we'd roll anything out.
Your constraints
Finally, the boring but critical stuff. What are your regulatory obligations? Does customer data have to stay in a particular country? What's your real budget, not the dream number? What systems do you already pay for that we should build on instead of replacing?
For a lot of clients, privacy is the live wire here. If you handle health records, financial data, or anything covered by privacy law, that narrows which tools are even allowed before we discuss features. Better to know on day one.
What you get at the end
The deliverable isn't a 90-page report you'll never read. We keep it to something you can act on:
- A ranked shortlist of two to four AI projects worth doing, with the highest-value one called out
- For each, a plain-language description of what it does, roughly what it costs, and what it would save
- An honest list of what needs fixing first, like a data source that has to be cleaned up
- A recommended starting point and a rough timeline
If the answer is "you're not ready," the report says so, and explains what would make you ready. We'd rather lose a project than sell you one that's going to fail.
How long it takes and what it costs
For a typical small or mid-sized business, an audit runs one to two weeks of our time, spread over a few weeks of your calendar so we're not living in your office. Most of the work on your side is giving us access and sitting through a handful of conversations.
The point of spending a little here is to avoid spending a lot in the wrong place. We've watched companies sink a five-figure sum into a tool that an afternoon of upfront questions would have ruled out.
A few things people get wrong about it
"We need to have our data perfect first." No. You need to understand your data. Perfecting it can happen alongside the first project, and the project often tells you which parts are worth perfecting.
"The audit will recommend the most expensive option." It usually does the opposite. The best first project is normally small and cheap, because that's how you learn with low risk.
"This is just for big companies." The smaller you are, the more a wrong technology bet hurts. The audit matters more when every dollar counts.
Is it worth it for you?
If you already know exactly what you want to build and your data's in good shape, you might not need a formal audit. Go build the small version and learn.
But if you're getting pulled in five directions by vendors, or you've got a sense that AI could help but no idea where, a couple of weeks of structured digging will save you months of expensive wandering. That's the whole pitch. No magic, just the questions someone should ask before you spend real money.
If you want to talk through whether your business is at that point, book a call. Even if the answer is "wait six months," you'll leave knowing why.