TECH AND MARKETING

Matt Goren, founder of RunOctopus, explains how artificial intelligence (AI) assistants recommend local businesses and what pest management companies can do to improve their chances of being found.

I should probably admit something up front. For years I sold cockroaches for a living. My company bred feeder insects for reptile keepers, so writing for people who get rid of them feels a bit like switching sides.

What I learned running that business is the reason I'm writing this. We were a small company in a market where the big names owned Google. We couldn't outspend them.

What we found was that when people started asking AI assistants for advice, the assistants didn't care much how big we were. They cared about who could actually answer the question, and who other people said good things about.

That's starting to matter for pest management too.

A homeowner finds droppings behind the cooker. A few years ago they'd have typed “pest control near me” and scrolled through a page of results.

Now a lot of them open ChatGPT or Google's AI answer and type a whole sentence, like “who's best for mice in a terraced house in Derby?” What comes back isn't a page of links. It's two or three names, with a reason next to each.

If you're one of those names, you get the call. If you're not, that customer probably never hears of you.

I tested it

Earlier this month I asked Google's AI Mode, ChatGPT and Claude the same everyday questions in two places in the US. One was Richmond, Virginia, a proper city. The other was DeLand, Florida, a small town.

I asked about coffee, pizza, air-conditioning repair, plumbers and oil changes. Not pest management, I know, but the same sort of job as yours: local, often urgent, and someone needs a person to call.

For plumbers in Richmond, the answers were led by local firms. The two big national plumbing brands only turned up in Google's sponsored adverts underneath. For coffee, pizza and air-conditioning repair, I didn't see a single national chain in any answer.

DeLand surprised me more. A plumber with 19 reviews was named in the answer, while firms with 4,600 and 10,400 reviews were sitting in the paid slots below it.

And when I looked at what the assistants were actually reading, it was nearly always other people talking about the business. Review sites, local forums, local news, trade listings. Claude said it straight out in one answer: what a company says about itself is “a lead, not an endorsement.”

To be fair to everyone, that was one day in two American towns, and the UK will have its own sources. But it matches what the AI companies say about how their systems work, and it's easy to check where you are.

It's a survey, really

You already know how this works, because you do a version of it every time you walk into a property.

You don't call it rats based on one dropping. You look at the droppings, the gnaw marks, the smears along the skirting, the runs in the loft insulation. Any one of those could mean something else.

When four of them agree, you know what you've got and where it's coming in.

The AI looks at a business the same way. Your own website is one sign, and on its own it doesn't prove much, because anyone can write anything about themselves. What tips it is a few independent sources agreeing with each other.

A review that mentions the actual job and the town. Your listing on a trade association directory. A mention in the local paper.

Someone on a community forum saying “we used them, they were good.” When those line up on your name, your area and what you do, the AI is confident enough to name you.

It also pays close attention to the exact question. “Wasps' nest in a loft in Leicester” isn't the same question as “pest control Leicester.” The more specific the customer gets, the more the answer leans towards firms that clearly do that job in that place.

So the big firm with the big budget can buy the advert next to the answer. Getting into the answer itself depends much more on what's written about you, and where.

The usual problems

The most common one I see is a single “Services” page that lists rats, mice, wasps, bed bugs, birds and moles in one go. Then someone asks about bed bugs in a flat in Nottingham, and there's nothing on the site that actually talks about that. The AI has nothing to quote, so it quotes someone else.

Close behind is the site that only mentions one town. You might cover 12, but if the website only names the one where you park the van, there's no reason for the AI to connect you with the other 11.

Mismatched listings cause more trouble than people think. An old phone number on one directory, the trading name on another, the limited company name on a third.

A person can tell it's the same firm. Software piecing things together isn't so sure, and when it isn't sure, it leaves you out.

Then there are reviews. “Great service, thanks” is lovely, and it helps your star rating. But “came out the same day for a wasps' nest in our loft in Belper and explained what he was using” is the one that matches you to a real question.

And most firms have simply never checked. It takes five minutes.

What I'd do this month

Think of it like placing bait boxes. You don't put one in the middle of the kitchen and hope for the best. You put them on the runs, where they actually travel.

Your information needs to be where the AI travels, and most of that is other people's websites, not yours.

1. Ask the AI yourself

Open ChatGPT and Google, in a private window if you can, and ask what your customers would ask, in their words. “Who can get rid of rats in my garden in [town]?” “Who does wasp nests in [town]?” Write down who comes up and which websites they're quoting. That's your real competition, and it might not be who you'd guess.

2. Give your main jobs their own pages

Rats, mice, wasps, bed bugs, birds, whatever you do most. On each one, answer what people actually ring up asking. How long will it take? Is it safe with the kids and the dog? What happens on the first visit? What do I need to do before you come? You answer these on the phone every day already, so write them down.

3. Write properly about the places you cover

A long list of town names at the bottom of the page doesn't do much. A short section on each main area, ideally with a real job you did there, does.

4. Make your listings match

Same name, number, address and area on your Google Business Profile, your trade association's member directory, and anywhere else you're listed. A listing with a recognised trade body is exactly the kind of outside source the AI trusts, so make sure yours is complete.

5. Ask happy customers to mention the job and the town when they leave a review

You're not telling them what to say, just reminding them what helps the next person.

6. Get other people talking about you

A quote in the local paper, a talk at a residents' association, a useful answer on a local Facebook group, a mention in a letting agent's newsletter. Each one is somebody else vouching for you, and that's what counts.

7. Ask the same questions again in a month

Treat it like a follow-up visit. What's changed since last time tells you whether it's working.

None of this needs a big budget. It mostly needs someone who knows the work to write about it plainly and make sure the rest of the internet agrees. In my experience, the people who know the work best are usually the small firms.

ABOUT THE AUTHOR

Matt Goren is the founder of RunOctopus, helping businesses get picked by ChatGPT and Google.

runoctopus.com

Source: