Look for a back office

The businesses worth approaching share one property: somebody sits at a desk doing repetitive work that customers never see.

A lettings agency processing referencing checks. A veterinary practice doing insurance claims. A builders' merchant matching delivery notes to invoices. A haulage firm doing driver paperwork. A recruitment agency formatting CVs.

Compare with a busy café. It has volume, energy and pain, and the pain is footfall, staffing and stock — none of which you can touch. A queue at a counter is not a queue you can automate.

The boring, unglamorous, back-office-heavy business is where this work lives.

They also have to be able to pay

This is uncomfortable and it saves months.

A business with the right shape and no money is a business that will enthusiastically take a free pilot and then never convert. You will have done the work, learned a lot, and be no closer to doing it again.

Signals worth reading: are they hiring? Do they have more than one location or more than about eight staff? Do they already pay for software beyond accounting? A business that has never bought software will find paying for it a much bigger step than the amount involved suggests.

The first engagement can absolutely be free or cheap — you are buying a case study. Just choose a business that could have paid, so the second conversation is possible.

Make a list of twenty, then cut it to five

Twenty local businesses of the right shape, from a map and twenty minutes. Then cut on two criteria: is there plausibly a back office, and could they plausibly pay.

Twenty gets cut to five fast, and the five are worth approaching properly rather than the twenty being approached badly.

Which of these would you approach?

Four local businesses, all within twenty minutes of you.

Candidates surviving the cut

How many of your twenty survived both filters. Note what killed the rest — it is usually the same thing twice, and that tells you where to look next time.

Before you knock on anything

You have a shortlist built on two unglamorous criteria rather than on what sounds interesting.