October 29, 2026 · Daniel Lawton
What an AI assistant says when someone asks for a plumber in Brockton, and how to be the answer
Type a question into a search engine today and, more and more often, you do not get ten blue links. You get a paragraph. For local services the paragraph tends to say something like: here are a few well-reviewed plumbers in Brockton, with two or three names, a sentence about each, and a phone number. Then it stops.
For the business that is named, that paragraph is the best advertising it never bought. For the businesses that are not, there is no page two to be on. This article explains how those answers get assembled, what the systems are actually reading, and the concrete things a South Shore business can do to become the answer.
What the assistant is doing
Whether it is Google’s AI summary, a chat assistant, or a voice assistant on a phone, the system is answering a question by reading sources it has access to and synthesizing what it finds. It is not ranking pages by secret formula so much as trying to produce a confident, specific, verifiable answer. That changes what matters.
It wants facts it can state plainly: this business does this service, in these towns, at these hours, at roughly this price, and here is evidence people were happy. It wants those facts to agree across the places it looks. It wants recency, because an answer about who to call today should rest on evidence from this year. And it avoids naming a business it cannot describe, because a vague recommendation is a bad answer.
A business whose website says we serve the greater Boston area, whose Google profile lists a different phone number from the site, whose last review is from 2023, and whose pricing is call for a quote gives the assistant nothing to work with. It will name someone else.
What it reads
The sources are not mysterious. In rough order of weight for a local service business:
Your Google Business Profile. The categories you chose, the services you listed with descriptions, your hours, your service area, your photos, and above all your reviews. The assistant reads review text, not just star counts, and it quotes specifics: fixed our furnace same day, honest about the price. A profile with fourteen reviews and no services listed is nearly invisible.
Your website. Especially pages that answer questions directly. A page titled emergency plumber in Brockton that says what you do, when you answer, what it costs, and what customers in Brockton have said is exactly what the system wants to cite. A homepage with a slider and a mission statement is not.
Structured data. Code on your pages that states facts in a machine-readable vocabulary: this is a local business with this name, address, and phone; these are its services; this is a question and this is its answer; this service costs between this and this. The assistant does not have to infer from prose; it can read the fact directly. This site carries structured data for the business, the founder, every published price, and every one of the hundreds of questions it answers, which is why it is cited for them.
Consistency across the web. Directory listings, social profiles, the chamber of commerce page, local news. When every source agrees on the name, address, and phone, the system is confident it is describing one real business. When they disagree, it hedges or skips.
Recent, specific evidence. New reviews that mention the town and the service. Photos of recent work. A post about a job completed last week in Whitman. Systems weight recency because the user wants to know who is good now.
What it ignores
Keyword stuffing, which was never good and is now actively useless. Hidden text. Fifty town pages that are the same paragraph with the town name swapped; systems recognize templated content and discount it. Review bursts, forty five-star ratings in a week followed by silence, which read as what they are. Design flourishes, animation, and anything else that does not convey a verifiable fact.
In other words, the things that used to game search engines do not work on systems built to produce a trustworthy answer. What works is being a clear, specific, consistently described, well-evidenced business. That is good news for real businesses and bad news for the agencies that sold tricks.
The six things a South Shore business should do
1. Finish the Google profile. Every service listed with a plain description. The most specific primary category. Hours that are true. Photos added monthly. Every review answered. The Q&A section populated by you with the questions customers actually ask.
2. Build a page for each real service in each real town. Not fifty templates: ten real pages, written individually, for the services and towns that produce your best work, with the local specifics, the recent jobs, and the reviews from that town. The town pages on this site are the model.
3. Answer the actual questions, in writing, on your own site. What does it cost. How fast do you come. Do you work weekends. Are you licensed and insured. Each question answered plainly, marked up as a question and answer in structured data so the assistant can lift it directly.
4. Publish your pricing, or at least your ranges. The single most common question about any service is what it costs. A business that states a range is the one the assistant quotes when the user asks. Fear of scaring off customers is misplaced; the customers you scare off were never going to pay your price, and the ones who stay arrive pre-qualified.
5. Make your facts agree everywhere. One name, one address, one phone, one set of hours, on your site, your profile, every directory, and every social account. Fix the old listing with the previous phone number. The system is checking.
6. Generate steady, specific evidence. A review request after every job, automatically. Photos of finished work. A short post naming the town. The Proof Engine in ICS’s systems does this without the owner touching a keyboard, because consistency is what the system rewards and consistency is what busy owners cannot sustain by hand.
What this looks like done
A Stoughton electrical contractor came to us with a decent reputation and almost no presence: a one-page site, a Google profile with nineteen reviews and no services, and pricing by phone only. We built service pages for the towns that produced his best work, answered forty real questions in writing with structured data, published his pricing ranges, aligned every listing, and turned on automated review requests. Within a quarter, AI-generated search summaries for electrician queries in his towns began naming him, with review language quoted. His phone did not ring more often; it rang with better jobs, from people who already knew his price.
Where to start
The AI Readiness Score on this site is a free, two-minute assessment that checks the things above and tells a business where it stands and what to fix first. The Ask ICS box in the site’s search is a live demonstration of the principle: ask it what Lead Engine costs or whether ICS is licensed in Florida, and it answers from the same structured facts an assistant would read.
The paragraph that names three businesses is where local search is going. The work to be in it is not a trick. It is being a clear, honest, well-documented business and making sure the machines can tell.
Common questions
- How do AI assistants decide which local businesses to recommend?
- They synthesize what they can read and verify: the business's own website, its Google Business Profile, reviews and their specific language, directory listings, local news, and structured data that states facts plainly. Consistency across those sources, specificity about services and towns, and recent evidence of real work weigh heavily. Vague, inconsistent, or thin presences get skipped.
- Is answer engine optimization different from SEO?
- It builds on SEO and changes the goal. SEO earns a position in a list; AEO earns a mention in an answer. The fundamentals overlap, but AEO puts more weight on machine-readable facts, explicit answers to real questions, published pricing, and a consistent identity the system can trust.
- Does publishing prices help a business get recommended by AI?
- Yes. When a user asks what something costs, an assistant prefers a source that states a number over one that says call for a quote. ICS publishes its own pricing as structured data for exactly this reason, and we build the same into client sites.
- How long does it take for AI search to start mentioning a business?
- Changes to a website and Google profile are typically re-read within weeks. Building the review base and the local evidence that makes a recommendation confident takes a few months. ICS's ninety-day guarantee covers measurable visibility, and the AI Readiness Score on this site shows where a business stands today.