“Use AI” is one of the fastest-growing searches in our data this year — up 60% in the UK. But before reaching for a new AI tool, it’s worth asking three questions:

1. Is the problem repetitive? AI is best at tasks you (or your team) do over and over: writing similar emails, categorising support tickets, summarising calls or documents. If a task is a one-off, AI may not save you much.

2. Is “good enough” actually good enough? AI output is rarely perfect first time. For low-stakes tasks (first drafts, internal notes, research summaries) that’s fine. For anything customer-facing or compliance-related, you’ll still need a human review step.

3. Do you have the data or process to plug it into? AI tools work best when there’s a clear input and output — a document to summarise, a ticket to triage, a list to clean up. If your process is undocumented or inconsistent, fix that first.

If your problem ticks these boxes, AI can genuinely save hours a week. If it doesn’t, the tool itself usually isn’t the bottleneck — the process is.