Definition
AI readiness is an assessment of preconditions rather than ambition. It asks whether your processes are defined, your data is reachable, your risk tolerance is understood, and someone is accountable for the outcome, because AI projects fail on those grounds far more often than on technical ones.
Readiness is not enthusiasm
Most AI readiness assessments measure the wrong thing: appetite, budget, and executive sponsorship. Those determine whether a project starts. They have almost no bearing on whether it works.
What determines that is far more mundane. Can a system reach the data at the moment it is needed? Is the process defined well enough that "correct" has a meaning? Does anyone own the outcome after the project ends? A business with enormous enthusiasm and none of these will produce an impressive pilot and nothing else, which is roughly the story behind the widely cited finding that around 95% of enterprise generative AI pilots produce no measurable return.
The four conditions that actually matter
Process clarity. Can you describe the target workflow end to end without saying "it depends on who's doing it"? If not, you have a process mapping problem first, and AI would only industrialize the ambiguity.
Data reachability. Not whether the data exists, but whether a system can query it now. Information locked in PDFs, individual inboxes, or a database nobody has credentials for is not available, regardless of how much of it there is.
Risk clarity. Do you know what a wrong output costs, how quickly you would notice, and who is accountable? Without this you cannot decide how much autonomy to grant, so you will either over-restrict the system into uselessness or under-restrict it into trouble.
Ownership. A named person responsible for the outcome, not just the project. Systems without owners drift, and nobody notices when they stop working properly.
Three of four solid is enough to scope something real. Two or fewer means the operational work comes first, and doing it produces value with or without AI.
What readiness is not
It is not having clean data. Perfectly clean data is a condition no operating business has ever met, and waiting for it is a way of never starting. You need reachable, reasonably reliable data for the specific process in question, which is a much smaller requirement.
It is not having a strategy document. It is not having chosen a model or a vendor, which is one of the last decisions rather than one of the first. And it is not scale: a 30-person business with one well-defined, high-volume workflow is readier than a 500-person business whose processes are all bespoke.
How Automathing approaches it
We assess readiness per process rather than per company, because the answer differs sharply between two workflows in the same business. A single process with clear steps, reachable data, and an owner is a viable project even in an organization that is unready overall, and starting there produces the evidence that makes the next one easier.
Frequently asked questions
How do we know if we are ready for AI?
Pick one specific process and answer four questions honestly: can you describe it end to end, can a system reach the data it needs, do you know what a mistake costs, and is someone accountable for the result. Three solid answers means scope it. Fewer means fix the operating problem first, which is cheaper and pays back regardless.
Do we need to clean our data first?
You need the data for the specific process to be reachable and good enough, not clean in general. A comprehensive data-cleanup programme before any AI work is a multi-year project that usually stalls. Scope the data work to the process you are actually building, and let the result justify the next round.
What is the biggest blocker for mid-sized businesses?
Data access, consistently. Not data volume, not model choice, not budget. Information exists but sits in systems that cannot be queried, formats that cannot be parsed, or accounts nobody can grant access to. This is why AI projects so often turn into integration projects, and why an integration foundation pays for itself twice.
Should we hire someone before starting?
Usually not first. The initial work is operational: mapping a process, establishing what data is reachable, and deciding what a mistake costs. That is work your existing team can do, often better than an outside hire who does not know the business. Hire once you know what you are building and intend to run it long-term.