What an AI Readiness Audit actually finds

You are excited about AI, and you should be. You can see that it could take real cost out of the business and give your team back the hours that drain them, so the business can scale without adding headcount at the same rate. What is harder to see is where it fits in your business specifically, what it will cost to get there, and what you should do first.

That gap is what an AI Readiness Audit closes. This article sets out what a genuine audit produces, what it costs, how long it takes, what it tends to find, and how to tell which parts of your business are ready to build on now and which need a little groundwork first.

Why does most AI readiness advice fall flat for an SME?

Most of it is written for enterprises with their own IT departments, or it is a free quiz designed to sell you something.

When a founder goes looking for guidance on getting ready for AI, most of what they find comes from one of two places, and neither was built for an operating business of six to fifty people.

The first is the vendor framework. The large technology companies publish detailed readiness models built around multiple pillars and maturity tiers. They are serious pieces of work. They also assume a dedicated IT function, a data team, a security function, and a steering committee, none of which a founder-led SME has or needs. Reading one tells you how a large enterprise should think, it does not tell you what to do on Monday.

The second is the free assessment. A short online quiz returns a readiness score in a couple of minutes. It feels useful, but it is built to capture your details and book you a sales call, and it commits to nothing on the other side: no fixed price, no timeline, no deliverable you can hold. Between the two sits a third habit worth naming, the price-on-application page, where the cost of the work stays hidden until you are far enough into a conversation to feel committed.

The effect is that capable businesses either stall or spend badly. The fix is not more reading. It is a piece of work with a defined scope, a fixed price, and a finite end, which is what an audit is.

What does an AI Readiness Audit actually produce?

A defined set of deliverables across three stages, taking you from where you are now to a costed plan for what to build first.

A real audit produces a defined set of artefacts you can act on, not a presentation and a follow-up call. At VeloBridge the report is a defined set of artefacts, grouped into three stages.

The first stage maps where you are now. It covers:

• A technology stack audit with a gap analysis.
• An operational touchpoint map showing where your core workflows meet your systems.
• A data architecture and hygiene assessment covering whether your data can be reconciled across those systems.
• A baseline of any AI already in informal use across the business, measured against the Australian Guidance for AI Adoption.

The second stage finds the opportunity. It covers:

• A prioritised list of AI and automation use cases scored by effort and expected impact.
• A build-versus-buy and configure-versus-custom decision for each priority.
• An honest read on how much new process the team can realistically absorb.

The third stage sets the direction. It covers:

• A model of the projected impact on margins and team capacity.
• A phased delivery plan with clear go or no-go points and indicative investment.
• A short set of criteria naming the areas where the recommendation is to do the groundwork before building.

Three of these are deliberately concise. The governance baseline, the change-capacity read, and the do-the-groundwork-first criteria are judgements rather than workstreams, and their value is in the decision they inform, not their length. They are also the three that most providers leave out entirely.

What does it cost, and how long does it take?

A fixed fee of A$12,000 to A$18,000 plus GST, delivered over 10 business days, with both the price and the timeline published up front.

The fee is fixed within that band before any work starts. Where you sit in the band depends on the number of systems in use, the size of the team we need to interview, and the operational complexity we find while scoping. There are no hourly surprises and no open-ended meter.

Ten business days is a deliberate choice. Much less than that and the work stays shallow, a scan rather than an assessment. Much more and the audit stops being an audit and turns into the project it is supposed to scope.

We publish the price because a buyer deciding whether to commission an audit needs to know what it costs before they invest time in finding out. A fixed fee you can see, set apart from any later build, lets you treat the audit as a self-contained decision. You can take the report, act on it yourself, hand it to your own team, or take it to another partner, and it remains useful either way.

What does an audit like this tend to find?

The findings about data and systems are the ones every audit looks for. The ones about people are what actually decide how fast and how well you adopt.

Some of what an audit surfaces is predictable, and we find it in almost every engagement. The first is fragmented data. The same customer exists in the CRM, the finance system, and an operations spreadsheet as three separate records, with no shared identifier to tie them together, so producing a single reliable view of a customer, a job, or a month of revenue means someone exporting and reconciling it by hand. Informatica’s 2025 research put data quality and readiness as the most cited obstacle to AI, named by 43 per cent of organisations, which matches what we see on the ground.

The second is undocumented process. The way work actually gets done lives in the heads of a few experienced people rather than in anything you could hand to a new starter, let alone to a machine. This is a separate problem from fragmented data, and it needs a separate remedy, because a unified data layer does nothing for process knowledge that was never written down. One is an architecture problem and the other is a knowledge problem.

These two are the foundations, and we have written before that AI can only work with good data and clear processes. They are also the findings every audit looks for. The two that decide how fast and how well a business adopts AI are different, and they rarely appear on a readiness checklist at all.

The first is an awareness gap. Most teams do not know what current AI and automation can actually do, so processes carry on unquestioned as the way they have always been done. Part of this is a belief, left over from an earlier era of automation, that only repetitive and identical work can be handed to a machine, so anything involving judgement or variation gets written off. That belief is now wrong, and closing the distance between what a team assumes is possible and what actually is may be the most valuable conversation the audit starts.

The second, and the one that catches the most capable businesses, is the limit on how much change a team can absorb. Once a business decides to act, the constraint is rarely budget and rarely the technology. It is the rate at which people can take on new ways of working without the wheels coming off.

We saw this with a fast-moving services scale-up. For years their bottleneck had been delivery: every change to their systems took weeks to build, and the standing frustration was that technology could not keep pace with the business. AI-assisted delivery removed that bottleneck and exposed a different one. We could now build and deploy faster than the team could absorb. Staff kept following processes that had already been superseded, because the replacements had not yet been learned, and that broke things downstream. The team began to feel change fatigue, and the speed that was meant to help them had started to work against them.

The fix was not to build more slowly. It was to govern the rate of change on purpose. We introduced a change gateway with three tiers, sorted by the impact a change has on the people using it rather than by how much engineering it took. A minor change needs no retraining and ships quietly. A major one alters how someone does their job, so it carries communication and training before it lands. A new-process change adds leadership sign-off and, where customers are affected, customer communication. A two-line configuration change can sit in the top tier if it alters what a customer sees, because the tier is set by impact, not effort.

What matters in that story is not the framework. It is that the framework was deliberately light. Three tiers, not seven. A scale-up cannot carry heavy process and stay fast, so the governance has to be sized to the business it serves and built to grow as the business matures. An enterprise change methodology dropped onto a thirty-person company creates more drag than the problem it set out to solve. Governance that fits is the finding, and the tiers are simply how it looked for that client, for now.

How do you know which parts of your business are ready?

Every business should be adopting AI. The audit decides which processes are worth it first, and which areas need a little groundwork before they are built on.

The question is not whether your business should use AI. It should, and the sensible time to start is now. The useful question is narrower: which processes are worth doing first, and which areas need their foundations sorted before anything is built on them.

Worth is not a question of size. The old rule was that a task had to be repetitive and identical to justify automating it. That rule no longer holds. Variation is now often the reason to use an agent rather than a reason against it, because an agent can reason through the exceptions and judgement calls that a fixed rules-based tool never could. So a process earns its place on the list when it carries enough regular volume to free up real time, or when it involves the kind of judgement that has tied up skilled people precisely because no simple script could handle it. The one caveat worth stating is that the time you free has to be redirected to higher-value work, or the saving stays on paper.

Readiness, on the other hand, is about the groundwork. An area benefits from a little preparation first when there is no agreed definition of a good outcome to build toward, when the data feeding it cannot yet be trusted or reconciled, when the team is already carrying as much change as it can absorb, or when the whole process depends on one person and would not survive their absence.

There is one more, and it is the one most often left unsaid. Sometimes a team fears that AI is there to remove them. The businesses we work with do not see it that way. They use AI to give their people back the hours that admin steals, so they can spend more time on the work that needs a human. Not every business thinks like that, and that is fine, but where the fear is real it is itself a readiness issue, because a team that believes a tool exists to replace them will not give it the process knowledge it needs or adopt it once it ships. The answer is design and honesty: AI that recommends while a person keeps the decision, which is also what the Australian Guidance means by keeping a human in control, alongside a clear account of what is being automated and why.

How does this apply across a portfolio?

For PE and VC operators, one audit method applied across the portfolio turns AI readiness into something you can compare and govern.

If you are an operating partner looking across several portfolio companies, a single audit method applied consistently gives you a comparable read on where each business sits, where the common opportunities are, and which companies are ready to move first. It replaces a set of one-off conversations with a board-ready view built on the same criteria for every company, which is the same logic we have described for standardising change across a portfolio.

Where do you go from here?

Start with the audit, before the build. It is the least expensive way to make the next decision a confident one.

If your systems are straining as you grow, and you can feel that AI should be helping somewhere without being certain where, the audit is the place to begin. It turns a general sense that you should be doing something into a costed, prioritised plan you can act on, whether you build it with us or not.

The single rule worth holding to is this: commission the audit before you commission a build, because the audit is what tells you which build is the right one to fund.

You can read more about our AI Readiness Audit, or get in touch for a straight conversation about whether it is the right next step for your business. We will tell you if it is not.