A self-test for AI readiness: Five questions before you spend anything

The conversation around AI readiness tends to be vendor-driven — a checklist designed to make you feel like you are almost ready, and that buying their tool is the final step. This is not that. These five questions will tell you whether an AI investment will deliver returns — or whether it will add to the pile of tools that did not work.

Question 1: Can you describe your current process in five steps or fewer?

Automation requires a defined process. If you cannot describe your current workflow in five steps or fewer — with a clear trigger, a sequence of actions, and a defined outcome — then automating it will encode the confusion into the system. Vague processes produce vague automation.

Before you automate anything, you need to be able to describe exactly what happens now. Not what should happen. What actually happens, step by step, who does each step, and what tool they use to do it.

Question 2: Do you know how long your current process takes?

Without a baseline, you cannot measure improvement. If you do not know that your current lead entry process takes 15 minutes per lead, you cannot calculate the value of automating it. You cannot determine whether a R18,000 integration pays for itself in six months or three years.

Question 3: Is your data clean and consistent?

AI and automation systems work with data. If your data is inconsistent — names entered in different formats, phone numbers with and without country codes, CRM records with missing fields — the automation will break or produce bad outputs. Garbage in, garbage out is not a cliché. It is the single most common cause of automation failure in the first month.

This does not mean your data needs to be perfect before you start. It means you need to know how clean it is and factor data remediation into the implementation plan.

Question 4: Do you have decision-making authority?

Workflow integration requires real decisions — which tools to connect, what data to move, how to handle edge cases, whether to change how certain tasks are done. If the person leading the implementation does not have the authority to make these decisions, the project stalls every time a choice needs to be made.

In practice, this means the person commissioning the work needs to be the person who can say 'yes, change how we do this.' If that requires committee sign-off, the implementation will be slow, expensive, and frustrating for everyone involved.

Question 5: Is your team willing to change how they work?

This is the question that matters most — and the one most businesses gloss over. Automation does not just change the tools. It changes the workflow. Processes that were manual become automated. Tasks that lived in someone's head need to be documented and systematised. Ways of working that have been in place for years get replaced.

If key team members are resistant to this change — if there is a tacit agreement that the new system will be humoured but the old way will continue in parallel — the integration will fail regardless of how well it was built. Research on technology adoption consistently finds that user acceptance is the primary determinant of whether a new system delivers its intended benefits.

What to do with this

Use these questions as a filter before any AI investment conversation. Not to disqualify yourself, but to understand where the preparation gaps are. A good implementation partner should be asking you all five of these questions before proposing anything. If they are not, they are selling you their process, not solving your problem.