AI Readiness Assessment: The Three Checks That Decide If You Are Ready
June 8, 2026 - by Themba Mahlangu - 5 min read
Before you spend money on AI, the real question is whether your business is set up to get value from it, or whether the project will stall. The answer comes down to three things you can check yourself. A model can reach your data. You have one specific job you want it to do. Someone will own the result. If all three are true, you are ready to invest. If one is missing, fix that first, and you avoid paying for a project that was never going to work.
We built and ran Hyper, a platform that put AI to work for a few hundred small businesses and agencies, and these three checks are the pattern we saw separate the businesses that got value from the ones that gave up.
Check one: can a model reach your data?
Most of what makes a business "AI-ready" is mundane. Can a model actually reach the data it would need? Can it read your inbox, your ad accounts, your CRM, your sheets, your analytics? If your customer data lives in one person's head and a spreadsheet they email around, a model has nothing to work with, and no amount of model quality fixes that.
When we ran Hyper, the clearest divide was between businesses that had connected their everyday tools and those that had not. The connected ones got real value. The others got generic answers and drifted off.
This is the most common reason AI projects stall, and the wider data backs it. Gartner predicts that through 2026, organisations will abandon 60% of AI projects that are not supported by AI-ready data, and that most organisations either do not have, or are unsure they have, the data practices AI needs (Gartner, February 2025). The bottleneck is rarely the model. It is whether the model can see anything useful.
The good part is that this is the cheapest gap to close. You do not need a data warehouse or a governance committee. You need your existing tools connected through credentials that hold.
Check two: do you have one specific job for it?
Readiness is also about knowing what you want the AI to do. The businesses that got value from Hyper almost always arrived with one specific job they wanted done. The ones who came to "explore AI" and see what was possible mostly drifted off within a few weeks. The most common job we saw was email triage: sorting an inbox into what needs action, what is waiting, and what is noise, then drafting the replies. It is dull, it runs every day, and the value builds up.
So the readiness question is simple. Can you name the one job you want AI to do, say roughly how many hours a week it costs you now, and name who will own it? If you can, you are ready. If you cannot, hold off and spend the time working out which job is costing you the most. That is the first thing a real audit does.
Check three: will someone own it?
AI that nobody owns stops getting used. One named person has to be responsible for whether the workflow works, and has to use it often enough to notice when it breaks. Part of that is catching mistakes. A model will sometimes produce a number that is plainly wrong, and someone who owns the workflow and uses it daily catches that at once, before it lands in a report and does damage. McKinsey's 2025 survey points the same way: the organisations actually getting value from AI are far more likely than their peers to have senior leaders genuinely owning the work (McKinsey, 2025).
What the assessment gives you
Run the three checks honestly and you land in one of three places.
All three are true: data reachable, one job, an owner. You are ready to build. The work is connecting the tools and shipping that one workflow.
Data is reachable and you have an owner, but no clear job yet. You are ready for an audit. The job is to find the one workflow worth automating before you spend on the build.
Your data is not reachable. Fix that first. It is usually a short piece of plumbing, and it is the prerequisite for everything else. Spending on AI before this is paying for a model that cannot see your business.
A real assessment ends with one of those three answers and a specific next step. A maturity score ends with a number. That is the difference that matters when you are deciding where to put money.
What it costs
For a small business, the flat number is straightforward. An AI audit, which is the readiness assessment done properly and ending in a decision, starts at $3,000, and we credit that against the build if you go ahead. If it leads to work, the assessment is effectively free. If it tells you to wait, it is cheap insurance against a failed project. The first integration, connecting your tools and shipping one workflow, usually runs $5,000 to $15,000, depending on how many systems are involved and how clean the data is.
For a larger organisation the question is return. Pick the workflow that consumes the most expensive hours, measure what those hours cost now, and measure what they cost once the workflow is automated and owned. If you cannot point to the hours, you are back at check two, and not ready to spend yet.
How we run it
We run the audit before we build anything because it is where these three checks get made against your real systems instead of a questionnaire. We connect to your actual tools, find the one workflow worth automating, and tell you plainly when the answer is not yet. Saying not yet is part of the job.
Sources
Gartner, "Lack of AI-Ready Data Puts AI Projects at Risk" (26 February 2025)