Three questions before you automate anything
Before we build any WattKind product, we ask three questions. At work I ask a version of the same three. They came from my years leading automation and AI in large organisations, where I watched good tools get bought and then left half-used, and they work just as well for a five-person business deciding whether to automate something.
1. Does this happen often enough to matter?
Automation pays back in repetitions. A task you do twice a year is usually better served by a checklist in a shared folder. A task you do every day, or that three people do every week, is worth a proper look.
2. Does it still work on your busiest week?
Most setups are designed on a quiet afternoon and tested on a quiet afternoon. Picture the week when two people are away and a big job lands. If the setup needs someone to remember a special step, check a dashboard or fix a broken connection, it will fail exactly when you need it.
3. Who maintains it?
Every automation needs an owner: someone who notices when it stops working and knows how to change it. If the honest answer is "the busiest person in the business" or "the person who set it up and has since left", the design isn't finished yet. Make it simpler, or don't build it.
What usually wins
More often than people expect, the answer is a better template, a shared checklist or a setting nobody had switched on. AI has a job to do, not a starring role. When it is the right tool, these three questions are what keep it working a year later.
If you'd like these questions asked of your whole business, the Opportunity Sprint does that in two days and leaves you with a written brief: what to do first, what it will roughly take, and what not to do.