Your Google Business Profile Is Feeding the AI Assistants Too

Ask an AI assistant to find you a plumber nearby and you will usually get a tidy list of real businesses with ratings and addresses. It is worth understanding where that list comes from, because it is not a mystery and it is not a new ranking system you need to learn.

In most cases it is Google Maps data, retrieved through Google's Places API. The assistant is not crawling the web to form its own opinion of local plumbers. It is asking Google, then writing up the answer in conversational language.

Why that is good news

The local SEO industry has spent the last two years selling anxiety about AI platforms: a new front door opens, and you supposedly need a new optimisation strategy for each one. For local business results, that is mostly not true.

If the assistant is reading the Places API, then the thing that determines whether you appear is your Google Business Profile, the same asset that determines whether you appear in the map pack. One asset, many front doors.

This is not a fringe theory. Google's own AI optimization guide names Business Profiles as the route for a local business to surface inside AI responses. Two different parts of the ecosystem pointing at the same file.

The part that is genuinely new

The API is public. Any developer can build on it, and with AI coding tools the cost of building a small local-search app has collapsed. Directory sites, comparison tools, niche apps, internal tools at bigger companies: a lot of them now pull live Google Maps data rather than maintaining their own listings.

The practical consequence: your profile is being read by software you will never see, in contexts you cannot audit, with no analytics coming back. You cannot optimise for each one. You can only make sure the underlying record is accurate and complete.

It also means errors propagate further and faster than they used to. Wrong hours are no longer a Google Maps problem. They are wrong in every downstream surface simultaneously, and none of those surfaces will email you about it.

What actually matters, in order

Because the data flows from a structured record, structured fields win. A beautifully written website paragraph is invisible to an API call. These are not:

• Primary category: the single strongest filter. Most API queries narrow by category before anything else.
• The services list: this is how a request for a specific job matches you rather than a generic competitor. It is also the field we most often find completely empty.
• Review count and rating: almost always what the assistant sorts and summarises by.
• Hours and service area: the "open now" and "near me" filters run on these.
• Attributes: the small tick-boxes. They are the only way a machine knows you do emergency call-outs.

How empty these fields actually are

We audited 1,440 Google Business Profiles: the top-10 map results across 29 US cities in five trades, collected in May and July 2026. The services section, the field that tells an API exactly what jobs you do, was unpopulated in 1,426 of those 1,440 profiles. That is not a rounding error. It is close to a universal gap among businesses that are already ranking well enough to appear in the top ten.

The pattern held across every trade we looked at. In roofing, 288 of 290 profiles had no services section filled in, against a median review count of 172. In HVAC, 287 of 290 were empty despite a median of 930 reviews, meaning businesses with substantial review histories had still never touched that field. Plumbing showed the same: 287 of 290 empty, median 728 reviews. Electrical came in at 285 of 290 empty with a median of 134 reviews. Landscaping was 279 of 280 empty, median 80 reviews.

The implication is straightforward. Review volume gets these businesses into the top ten. But the structured data a downstream API would use to match them to a specific job request ("replace a flat roof," "install a smart thermostat," "run a new circuit for an EV charger") is simply absent for nearly all of them.

The other gaps compound the problem. In the same sample, 65 percent of profiles had no booking link, every single profile had gone more than 90 days without a post, and 93 percent had no Q&A entries at all. Four percent had no website linked. These are not obscure optional fields. They are the signals that distinguish a complete record from a stub.

What to do about it: a specific sequence

The following steps are ordered by impact, not by how long they take. Start at the top and work down.

1. Verify your primary category. Log into your Business Profile dashboard and confirm the primary category is the most specific accurate description of what you do. "Roofing contractor" outperforms "contractor" in category-filtered API queries. If your primary category is too broad, everything else you do here matters less.
2. Build out the services section completely. This is the highest-leverage action for most businesses, because the gap is so nearly universal. Go to the Services tab and add every distinct job type you actually take on. Be specific: "flat roof replacement" and "roof leak repair" are different services and should be listed separately. Write a short, accurate description for each one. This is the field an API uses to match a specific customer query to your profile. Leaving it empty means you are relying entirely on category and reviews to match, which is a much blunter instrument.
3. Audit your hours for accuracy, including special hours. Set holiday hours and any seasonal variations before they happen, not after. An "open now" mismatch is one of the fastest ways to lose a referral from an AI assistant, because the assistant will state your hours as fact.
4. Check and set your service area. If you travel to customers rather than having them come to you, define the cities or zip codes you actually serve. The "near me" filter in most API queries runs against this field. A blank service area defaults to a small radius around your pin, which may exclude legitimate jobs.
5. Work through the attributes list. Attributes are the tick-boxes inside the profile editor that cover things like emergency availability, online estimates, or whether you serve commercial clients. They are the only structured signal for many of these characteristics. A customer asking an AI assistant for an emergency electrician available tonight is filtered in part on attributes, not on review text.
6. Add a booking link if you have one. This is not primarily an AI optimisation step. It reduces friction for any customer who finds you through any channel. But it is a populated field versus an empty one, and complete profiles consistently outperform incomplete ones in ranking experiments.
7. Answer those questions somewhere that still exists. The Q&A section was removed in December 2025, so the answers belong in your service descriptions, your business description and your website. Write the questions your customers actually ask before calling: "Do you offer free estimates?" "What areas do you serve?" "Are you licensed and insured in [state]?" An AI assistant retrieving your profile can surface this content directly. Leaving the section empty means you have no control over what appears there, because anyone can post a question.
8. Post at least once a month. Every profile in our sample had gone more than 90 days without a post at the time of collection. A post does not need to be long. A photo of a completed job with two sentences describing the work and the location is enough. Freshness is a signal. A profile that has not been touched in three months reads as less reliable than one that was updated last week, both to ranking algorithms and to customers who see the last-post date.

Who this does not apply to

This entire argument rests on AI assistants retrieving local results through the Places API. That is accurate for the general-purpose assistants, the ones a homeowner uses when they say "find me a plumber near me." It is less applicable in a few specific situations.

If your business operates exclusively in a B2B context where procurement decisions run through an RFP process or a vendor database, your Google Business Profile is not the decision point, and optimising it will not change that. The same is true if your customers find you entirely through referral networks, trade relationships, or platforms like Houzz or Angi that maintain their own databases and do not pull from Google Maps.

It is also worth being direct about the limits of any one data set. Our sample covers 1,440 profiles across five trades and 29 cities. It is not a random sample of all US businesses. The trades we selected (roofing, HVAC, plumbing, electrical, and landscaping) are all high-competition residential service categories. The pattern we found may be less pronounced in lower-competition markets or in trades that attract more digitally active business owners.

Frequently asked questions

If I already rank in the top three on Google Maps, does any of this matter?

Ranking in the map pack and being well-matched to a specific API query are related but not identical. A profile can rank well on broad category searches while still being invisible to a query for a specific service, because the services section is empty and the API has no structured data to match against. Say you run an HVAC company and someone asks an AI assistant for a contractor who installs mini-split systems specifically. If that service is not listed in your profile, the API has no reliable way to surface you for that query even if you rank well overall.

Do I need to pay for a tool to do this?

No. Every field described in the steps above is editable directly inside the free Google Business Profile dashboard at business.google.com. A tool can be useful for auditing at scale or for identifying gaps you have missed, but none of the optimisation steps require paid software.

How often should I revisit the profile after the initial setup?

Check the core fields (category, services, hours, service area) once a quarter and whenever your business changes. Attributes and Q&A can be reviewed less frequently, perhaps twice a year. The monthly post cadence is the only task that needs to be built into a regular routine. Google can and does suggest edits to your profile based on user submissions and its own data; you will receive notifications when this happens, and it is worth reviewing those promptly rather than letting an automated suggestion stand unchecked.

Will filling in the services section hurt my ranking if I list too many services?

There is no public evidence that a longer services list creates a penalty. The risk of over-listing is practical rather than algorithmic: if you list services you do not actually perform and a customer contacts you for one of them, that creates a friction point that can lead to a negative review. List what you genuinely offer. The goal is accuracy, not volume.

What to do about it

Nothing exotic. Fill in the structured record properly, keep it current, and stop paying for AI visibility products that promise to optimise you for individual assistants. There is no separate lever there to pull.

The honest summary is that AI has not created a new discipline for local businesses. It has raised the cost of a badly maintained profile, because the same record is now answering in more places at once.