A growing share of homeowners now ask an assistant before they open a search results page. That does not require a new marketing budget, but it does change which details on your site and profile matter — and most contractors are invisible for reasons that take a week to fix.
In this article
- What actually changed in local search
- How an AI answer gets assembled
- Citations are the new rankings
- Your business profile is the foundation
- Consistency beats volume
- Write the questions your customers actually ask
- Put the answer in the first two sentences
- Publish price ranges, with the conditions
- Structured data does the translating
- Reviews are read as evidence, not decoration
- Recency is a filter
- Being mentioned elsewhere still matters most
- Say exactly where you work
- What this looks like trade by trade
- The website still has a job
- Speed to answer decides the ones that arrive
- How to know whether any of this works
- What this is worth in a month
- The mistakes to avoid
- Where to start
- Frequently asked questions
- Frequently asked questions
What actually changed in local search
For fifteen years, getting found meant one thing: rank on the first page of Google for "roofer near me" and answer the phone. That path still produces work, but a growing share of homeowners now start somewhere else. They ask an assistant. They read an AI summary sitting above the map pack. They type a full sentence instead of two words. The result is that a contractor can hold the same ranking as last year and still watch calls decline, because the click never happened — the answer was delivered before anyone reached a website.
This is not a reason to panic or to buy anything new. Most of what makes a company appear in AI answers is the same material that makes it rank well in ordinary search, arranged slightly differently. What changes is which details matter and how explicitly they must be written down.
How an AI answer gets assembled
When someone asks an assistant to recommend a contractor in a city, the system does not consult a secret directory. It pulls from sources it can read and verify: business profiles, review platforms, directory listings, local publications and the contractor's own website. It then assembles a short answer with two or three names and a sentence about each. Being included depends on two things — existing in enough consistent sources to be considered real, and having facts written plainly enough to be quoted.
That second half is where most contractors lose. A site that says "quality workmanship since day one" gives the model nothing. A site that says "licensed and insured septic contractor serving Hamilton County, emergency service available seven days, typical tank replacement $6,500 to $11,000" gives it four quotable facts.
Keep reading — free
The rest of this article is worth the 20 seconds
Tell us where to send it and the full piece unlocks right here, along with everything else on this blog.
Rather just talk? Message us on WhatsApp.
Citations are the new rankings
The useful metric is no longer only position. It is whether your business is mentioned when the question is asked. A company can be cited in an AI answer without ranking first in traditional results, and a company can rank first and never be cited. Both happen every day. The practical consequence: consistency of your information across the web now matters as much as the strength of any single page.
Your business profile is the foundation
Every AI answer about local services leans on structured business data, and the Google Business Profile remains the densest source of it. Complete every field: categories, service areas, hours including holiday hours, services with descriptions, payment methods, attributes and photos with recent dates. Half-filled profiles get skipped in favor of complete ones, because a complete profile reduces uncertainty for the system doing the summarizing.
Consistency beats volume
Name, address and phone number should match character for character everywhere they appear: profile, website footer, directories, review sites, invoices. A company listed as "Smith & Sons Plumbing" in one place and "Smith and Sons Plumbing LLC" in another, with two phone numbers, reads as two weaker businesses instead of one strong one. Fixing this is tedious and free, and it is the single highest-return hour most contractors have available.
Write the questions your customers actually ask
AI search is conversational, so the queries are longer and more specific: "how much does it cost to replace a water heater in Denver", "do I need a permit to rebuild a deck", "what is the difference between a repair and a reline". Each of those is a page you can write. The old approach of one thin page per service no longer covers the range of questions a homeowner asks on the way to hiring someone.
Put the answer in the first two sentences
Pages that get quoted share a format: the direct answer first, the detail after. If the question is what a service costs, the first sentence contains a range and what moves it. If the question is whether a permit is required, the first sentence says yes or no and names the condition. Long introductions that delay the answer get skipped. This format also helps human readers, which is why it costs nothing to adopt.
Publish price ranges, with the conditions
Most contractors refuse to publish prices, and this is now expensive. Cost is the most common question asked of assistants, and an answer will be assembled from whoever did publish — often a national aggregator with numbers unrelated to your market. Publishing a range with the variables that move it ("$4,200 to $9,800 depending on roof pitch, tear-off layers and material") does not commit you to a number. It makes you the source instead of the alternative.
Structured data does the translating
Schema markup — LocalBusiness, Service, FAQPage, AggregateRating where legitimate — turns your page into machine-readable facts about what you do, where, at what price and with what rating. It is invisible to visitors and it removes guesswork for everything that reads your site. Any competent web developer adds it in an afternoon, and most contractor sites still do not have it.
Reviews are read as evidence, not decoration
Review text is quoted constantly in AI answers, and the text matters more than the star count. Reviews that mention the specific service, the neighborhood and the outcome create quotable evidence: "replaced our sump pump in Westfield the same day we called" does more than "great job, highly recommend". When you request a review, ask the customer to mention what was done and where. That single change in your request message reshapes what gets said about you.
Recency is a filter
Systems summarizing local businesses lean toward fresh signals. A profile with photos from two years ago and a most recent review from last spring looks dormant next to a competitor posting monthly. The remedy is small and ongoing: a few job photos each month, a handful of recent reviews, updated hours and service descriptions when they change.
Being mentioned elsewhere still matters most
What others say about you outweighs what you say about yourself. Supplier and manufacturer dealer listings, chamber of commerce pages, local news mentions, trade association directories, sponsorship pages from a sports league you support, and partner pages from adjacent trades all count. Ten of those produce more effect than a hundred generic directory submissions, because they are the sources a system trusts when deciding whether a business is real and locally established.
Say exactly where you work
Vague coverage claims hurt. "Serving the greater metro area" cannot be matched to a question about a specific suburb. Name the cities, the counties and the neighborhoods you serve, and say which ones carry a trip charge. Location pages are only useful when they contain genuinely local content — the permit office you deal with, the common housing stock, a project you completed there. Twelve near-identical city pages with the name swapped are recognized as thin and ignored.
What this looks like trade by trade
The questions homeowners ask an assistant differ by trade, and so do the pages worth writing. A roofer is asked whether insurance will pay for a replacement, how to tell hail damage from wear, and what a square costs installed. A plumber is asked whether a water heater should be repaired or replaced, what a main line camera inspection costs, and who is responsible for the line at the street. An HVAC contractor is asked what size system a house needs, what SEER rating is worth paying for, and whether a heat pump works in local winters. A remodeler is asked for cost per square foot by room, how long a kitchen takes start to finish, and whether permits are required for a layout change.
Write those. Each one is a page with the answer in the opening lines, a range where money is involved, and a local detail that no national site can supply — the inspector's typical turnaround in your county, the housing stock built in a particular decade, the supplier lead time you deal with. That local detail is what makes your page the better source when a system is choosing between yours and a generic article, and it is the part competitors almost never bother to add.
The website still has a job
If an assistant recommends you, the homeowner still lands on your site to confirm the decision. That visit is short and it is looking for four things: proof you are licensed and insured, evidence of work like yours, an obvious way to contact you, and a signal that you respond quickly. A site that buries the phone number, loads slowly on a phone or shows stock photography undoes the recommendation that brought the visitor.
Speed to answer decides the ones that arrive
AI search tends to shorten the consideration phase. The homeowner gets two or three names at once, and contacts them in order. In practice this makes response time even more decisive than before, because the list is short and the second call happens within minutes. Answering live, or returning a missed call with a text inside five minutes, converts a meaningful share of inquiries that competitors lose by the end of the hour.
How to know whether any of this works
Three measurements are enough. First, ask the main assistants the five questions your customers ask, once a month, and record whether you appear — it takes fifteen minutes and is the only direct read available. Second, watch referral traffic from AI sources in analytics, which now appears as its own set of referrers. Third, and most reliable, ask every caller how they found you and log the answer; the phrase "an AI recommended you" has gone from never to regular in a year.
What this is worth in a month
Take a contractor booking 38 jobs a month at an average ticket of $2,400, roughly $91,200. Suppose AI-assisted discovery now influences 12% of inquiries in that market, and the company currently appears in none of those answers. Capturing even half of that share adds about 2.3 jobs a month, near $5,500 in revenue, against a cost of a few structured-data tags, a rewritten pricing page and a changed review request message. The same work also improves ordinary rankings, which is why the expected return holds even if adoption in your market grows more slowly than predicted.
The mistakes to avoid
Do not pay for "AI optimization" services promising placement in AI answers — no such placement is sold. Do not publish machine-written pages in bulk; thin repetitive content is the easiest thing for these systems to discount. Do not invent awards, certifications or review counts, because inconsistency between sources is precisely what gets a business dropped. And do not abandon the fundamentals that still produce most of the phone calls: profile, reviews, response speed and a fast site.
Where to start
Pick one week. Day one, fix name, address and phone everywhere and complete the business profile. Day two, write the five most common customer questions as five pages with the answer in the first two sentences. Day three, publish price ranges with conditions. Day four, add structured data. Day five, change the review request to ask for service and neighborhood. That is the whole program, it costs nothing but attention, and it is the same work that would have improved traditional search results anyway.
Being recommended is the new version of ranking first.
We build the profile, page set, structured data and review process that get home service companies cited when a homeowner asks for a recommendation, and we report results in booked jobs.
Frequently asked questions
How do contractors get found in AI search?
By existing consistently across the sources these systems can read, and by writing facts plainly enough to be quoted. In practice that means a completely filled business profile, identical name, address and phone everywhere, reviews whose text names the service and the neighborhood, pages that answer specific customer questions in the first two sentences, published price ranges with the conditions that move them, structured data markup on the site, and mentions on third-party pages such as supplier listings, associations and local publications. Nothing in that list is bought and nothing is new: it is the same foundation that improves ordinary local search, arranged so a machine can quote it.
Can I pay to appear in AI answers?
No, and anyone selling placement in AI answers is selling something that does not exist. There is no paid inclusion in an assistant's recommendation the way there is in search advertising. Worse, two shortcuts actively hurt: publishing bulk machine-written pages, which is the easiest kind of content for these systems to discount, and inventing awards, certifications or review counts, since inconsistency between sources is exactly what causes a business to be dropped from an answer. The work that produces inclusion is unglamorous — profile completeness, data consistency, specific reviews and pages that answer real questions directly.
How do I know if AI search is sending me customers?
Three measurements are enough. Once a month, ask the main assistants the five questions your customers ask and record whether your company appears; it takes fifteen minutes and is the only direct read available. Watch referral traffic from AI sources in your analytics, where these referrers now appear as their own group. Most reliably, ask every caller how they found you and log the answer — the phrase describing an AI recommendation has gone from never heard to regular within a year in most residential trades. If you track nothing else, track the third one.