Once you decide to automate, the hardest question is where to begin. Try to fix everything at once and you stall. Pick the wrong starting point and you spend effort on something that barely moves the needle. The trick is to choose deliberately, using a simple test rather than a gut feeling.
The short answer for most UK SMEs is to start where time and money are leaking fastest, usually slow lead follow-up or repetitive admin. A focused AI workflow automation project on that one task proves the value quickly and funds whatever you build next.
Start with pain, not technology
The most common mistake is starting with a tool you saw and looking for somewhere to use it. Do the opposite. Begin with the task that frustrates you most or costs you most, then find the automation that fixes it. Technology should follow the problem, never lead it. A quick way to surface candidates is to ask your team which jobs they dread and which they keep putting off. Those answers point straight at the repetitive, low-value work that automation removes best.
Score each task
Make a short list of your most repetitive or painful tasks and score each one out of five on three things: how much time it eats, how much it costs when done late or badly, and how simple it is to automate. Add the scores. The task at the top is almost always your best first move, because it combines high value with low effort. Resist the urge to start with the most exciting idea. The job of the first automation is to deliver a clear, fast win that builds confidence and frees up time, which then funds the more ambitious work later.
The usual first winners
In practice, a few tasks rise to the top again and again. Instant lead response and automatic follow-up win for any business that lives on enquiries, because every slow reply is potential revenue lost. Getting enquiries off your website and into your CRM without retyping is another favourite. So is invoice chasing, which is repetitive, easy to put off, and directly tied to cash flow.
A simple worked example
Say a builder spends three hours a week chasing quotes and loses a couple of jobs a month to slow replies. That task scores high on time, high on cost and simple to automate. Compare it to tidying an internal spreadsheet, which eats less time and loses no money. The follow-up automation wins easily, and the result is visible in weeks rather than months.
Avoid these first-automation mistakes
Do not try to automate a process you cannot describe clearly, because automation only multiplies whatever is already there. Tidy the process first, then automate the tidy version. Do not start with something so sensitive that a single error is costly, and do not over-engineer the first build. A tool like Zapier is often enough to prove the idea before you invest in anything more advanced. Finally, do not try to automate everything at once. Momentum comes from finishing one workflow well, seeing the result, and using that confidence to tackle the next, rather than stalling on an ambitious system that is never quite ready.
Your next step
You do not have to make this call alone. If you are weighing two or three candidates and want a second opinion, an AI automation consultant can score them with you. Better still, book a free AI Workflow Leak Audit, and we will pinpoint the single automation that would save you the most time and protect the most revenue, so your first move is also your best one.