Where to actually start with AI (and what to ignore)
The noise is the hard part, not the technology. Here's the first, safest, highest-return place to begin — and what you can skip without guilt.
Published · 5 min read
You already know AI matters. That's not the problem. The problem is that every article, every advert, and every well-meaning friend is pointing you at something different — and the fear of betting on the wrong thing is quietly stopping you from betting on anything at all. Months go by. The work that AI could be doing keeps getting done the slow way, by people who have better things to do.
Here's the reassuring part: the right starting point is not a mystery, and it's almost never the flashy thing. Once you know what to look for, the decision gets simple.
The menu is the enemy, not the technology
Most businesses don't stall on AI because it's too hard. They stall because the options feel infinite. When everything is possible, nothing gets chosen — and the cost of that indecision never shows up on an invoice, so it never gets fixed.
The way out is to flip the question. Don't ask "what can AI do?" — the honest answer is "far too much to be useful." Ask instead: "what does my business do over and over that nobody enjoys?" That question has a short answer in every company we've ever spoken to. And that short answer is your starting point.
What you can safely ignore
Permission to skip things is half the value of good advice, so here it is. For a first project, you can ignore:
- Anything pitched as a "transformation." Rebuilding the whole business around AI is a year-two conversation, not a week-one one. Start smaller and the transformation happens on its own.
- The demo-day showstoppers. The most impressive-looking AI is rarely the most profitable. The profitable kind is usually boring to watch and wonderful to own.
- Tools your business doesn't have a problem for. If you're adopting something because it's everywhere, not because it removes a specific pain, it will join the pile of subscriptions nobody opens.
None of this means those things are worthless — it means they're not first. First should be safe, fast, and measurable.
The best first move: the boring, repeated middle
The highest-return starting point in almost every business is the same: the repetitive, rules-ish work that sits between "a customer did something" and "the job is done." Reading incoming emails and sorting them. Retyping details from one system into another. Drafting the same routine replies. Chasing documents. Assembling the Monday numbers.
This work is ideal for a first AI project for three reasons. It happens constantly, so even a modest improvement compounds every single day. It's low-risk, because a human can review the output while trust is being earned. And it's measurable — you can count the hours before and after, so you're never guessing whether it worked.
As an illustrative example only: a task that takes a team forty minutes a day, automated down to five, returns hundreds of hours a year — from one unglamorous fix. That's the shape of a good first win: small on the surface, large underneath.
How to recognise your version of it
A strong first candidate usually ticks these boxes: it happens weekly or daily; it follows a pattern a patient person could write down; a mistake in it is annoying rather than catastrophic; and your team audibly groans about it. If a task ticks all four, you've likely found your starting point — no strategy retreat required.
What to do about it
- List the groan-work. Ask your team which tasks they'd pay to never do again. Write down the top three.
- Pick the one with a paper trail. Work that lives in emails, documents, and spreadsheets is the easiest to hand to AI well.
- Start with a small proof-of-concept. Prove it on your real work, with a human checking the output, before you commit to anything bigger.
- Measure one number. Hours saved, response time, or backlog cleared. One honest number beats ten vanity metrics.
Do this and the paralysis disappears — because you're no longer choosing from an infinite menu. You're solving one known problem, safely, with a visible payoff.
If you'd rather not work it out alone, our team does this every week. Book a $150 consultation and we'll help you find your best first project — and tell you honestly if something isn't worth doing yet. Book your consultation and start with the sure thing, not the shiny thing.