This is the question I get in almost every first conversation with a business owner.
They've heard the pitch. They know AI can automate things. They're interested. And then they ask: "But how do I know where to start?"
Most of the answers floating around online are useless. "Automate repetitive tasks." "Look for bottlenecks." "Find low-value work." Thanks. Very helpful.
Here's how we actually think about it.
Ask the Right Question First
The wrong question is: "What can I automate?"
The right question is: "What is costing me the most time right now, and how much judgment does it actually take?"
Those are two different questions, and the second one gives you an answer you can act on.
Every task in your business sits somewhere on a spectrum. On one end: high repetition, low judgment. On the other: low repetition, high judgment. The sweet spot for automation is obvious once you frame it that way. High repetition, low judgment. That's your starting list.
The Three-Part Filter
When we work with a client, we run every candidate process through three questions:
- How often does it happen?
Daily beats weekly. Weekly beats monthly. If something happens once a quarter, it's probably not your first target, even if it's annoying. - How long does it take?
An hour a day is 250 hours a year. Even shaving that to 20 minutes gets you back 167 hours. That's a month of work time. If you want to see what that math looks like across your whole team, our ROI Calculator (https://www.metroaipartners.com/roi) runs the numbers in about 30 seconds. - Does it require a human judgment call, or just consistent execution?
Sending a follow-up email to every new inquiry: no judgment required. Deciding how to respond to a complicated client situation: judgment required. These are not the same thing, even though they both involve email.
Run your candidate processes through those three questions. The ones that score high on frequency and time, and low on judgment, go to the top of the list.
What This Looks Like in Practice
A Government Contracting firm we worked with had a proposal coordinator spending 90 minutes every Monday pulling together a status update on active bids, gathering updates from three different people, formatting them, sending the summary to leadership. Same format every week. Same sources. Same output.
Judgment required: almost none.
Time cost: 90 minutes times 50 weeks equals 75 hours a year.
That's the first one you automate. Not because it's glamorous, but because it's undeniably worth it and the risk of getting it wrong is low.
A real estate office we worked with had the same pattern. Administrative reporting, lead follow-up sequences, showing confirmations. Things that happened constantly, followed a predictable pattern, and didn't need a human to decide anything. Those went first.
The Mistake Most People Make
They try to automate the complicated stuff first.
They want to build the big system. The AI that handles customer escalations, the tool that writes custom proposals from scratch, the workflow that manages exceptions. Those are real opportunities. But they're not your first move.
Start with the boring stuff. The task that happens every day that you or someone on your team does exactly the same way every time. Automate that. Build confidence in the process. Then move up the complexity ladder.
The businesses that get the most out of AI don't do it all at once. They start with one well-chosen process, see it work, and then keep going.
A Simple Rule to Walk Away With
If you can explain the task to a smart new hire in under ten minutes and they'd do it the same way every time, AI can probably do it. That's your filter.
Write down everything you or your team does in a week. Circle the ones that fit. That's your automation roadmap.
You don't need a consultant to build a 40-slide deck to figure this out. You need an honest list and a starting point.
Metro AI Partners trains and helps organizations build that list and actually execute on it. In-person, live on Zoom, or through our on-demand training library.
Start at metroaipartners.com