When is your business actually ready to automate a process?

A business is ready to automate a process when that process is repetitive, rule-based, high-volume, and stable - and when someone owns the outcome it affects. If a task is done often, follows steps you could write down, and has not changed much in months, it is a strong candidate. If it changes constantly or nobody agrees how it should work, fix the process before you automate it. Automating a broken process just makes the mess happen faster.

You do not need an AI strategy, a data warehouse, or a technical team to start. You need one good candidate.

What makes a process worth automating?

Run any task through these five questions. The more "yes" answers, the readier it is:

• Is it repetitive? The same task, done the same way, again and again.
• Is it rule-based? You could write the steps down and a sensible person would follow them identically.
• Is it high-volume or time-consuming? It happens often, or each instance eats real time.
• Is it stable? The process has not changed much recently and is not about to.
• Is the data reasonably structured? The inputs live somewhere you can get at them.

Tasks that score well are the usual suspects: CRM updates, invoice and document processing, month-end close, onboarding, report generation, ticket routing. These are exactly the process automation projects that pay back fastest - and note what they have in common: the steps are knowable in advance, so they want a workflow rather than an autonomous agent.

What are the signs you are NOT ready?

Be honest about these, because they are where automation projects fail:

• The process changes constantly. Automating a moving target wastes money. Stabilise it first.
• Nobody agrees how it actually works. If three people describe the process three different ways, you have a process problem, not an automation problem.
• The data is a mess. If the inputs are inconsistent or scattered with no system of record, fix the source first.
• There is no owner. If no one owns the outcome, no one will maintain the automation or act on what it frees up.

None of these are permanent blockers. They are just work to do first - and often the discovery itself surfaces them.

Do you need an AI strategy first?

No - and waiting for one is the most common way businesses stall. You do not need a grand strategy, new infrastructure, or a hire to automate your first process. You need a single high-value candidate, proven quickly. The ROI from that first win is what unlocks budget and appetite for the next one. Strategy is better written after you have shipped something real than before.

How do you pick the first one?

When several processes qualify, pick the one with the best ratio of impact to effort: high time or cost saved, low complexity to build. A deliberately small first project that pays for itself in weeks beats an ambitious one that drags on. Prove the model, bank the saving, then scale - that is the whole approach behind working AI automation that pays for itself.

Once you have your candidate, the next question is how to run it without it turning into an open-ended science project. The answer is a time-boxed pilot with a number attached to it: how to run an AI pilot that pays for itself in 60 days.

The bottom line

If you can name a task that is repetitive, rule-based, done often, stable, and owned by someone who cares about the result - you are ready, today. If you cannot, the first job is not automation; it is tidying up the process. Automating a broken process only makes the mess happen faster, and no amount of AI fixes a step nobody agrees on.

Either way, a thirty-minute discovery call will tell you which camp you are in - and I will say so plainly if the honest answer is "fix the process first". See process automation and AI consulting for how I approach the work, and the case studies for what a good first candidate has been worth.