"Near me" in 2026: how homeowners ask AI for trades, and what that means for you

From keywords to questions

For years, the way to find a trade was a short search: "furnace repair near me", "plumber Mississauga". Google reported back in 2016 that mobile "near me" searches had grown 136% in a year in the US. The phrase became a marketing obsession.

In 2026, that kind of search still happens, and the map pack still answers it. But there's now a second way to ask, and it looks different:

• "My furnace keeps clicking but won't light. It's a 2011 Carrier. Who in Scarborough can come today?"
• "Is it worth replacing a 20-year-old AC with a heat pump in Toronto? Who installs them?"
• "My basement floor drain is backing up after the rain, what should I do and who should I call in Etobicoke?"

These go to ChatGPT, Gemini, or Google's AI Mode, which started rolling out in Canada in August 2025 and is designed for longer questions with follow-ups. Nobody types "near me" in them, and they don't need to.

The AI already knows roughly where they are

• ChatGPT: OpenAI's help centre says ChatGPT search uses your approximate location, based on your IP address, to give relevant results, and may share a general location with trusted third-party providers for things like nearby business listings.
• Gemini: Google's help pages say Gemini uses public information from Google Maps to find places and show details like hours and ratings.
• Google Search and AI Mode: local results are based on relevance, distance and prominence. Distance is worked out from the searcher's location when they don't name a place.

So "near me" is now mostly implied. Whether the homeowner types it, names their city, or says nothing, the engine is trying to find a business that is close, relevant to the problem, and well regarded.

What changes for you

1. The answer is a shortlist, not a page of links

A typical AI answer to "who should I call" names a few businesses, with a line about each. Being sixth on a page of results used to mean some clicks. Not being in a three-name answer means none from that answer. That raises the value of the basics: being clearly relevant, clearly local, and clearly trusted.

2. Questions are about problems, not services

Homeowners describe symptoms: clicking, no heat, water on the floor, a burning smell. If your website only says "Furnace services", the engine has to guess that you handle ignition problems. If you have a furnace repair page that lists "furnace clicks but won't ignite" as a sign you need repair, there's a direct match.

Do this: on each service page, list the symptoms customers describe when they call. Our service page template has a section for exactly this.

3. Specific details decide the shortlist

Longer questions carry conditions: "today", "2011 Carrier", "heat pump", "after the rain", "Scarborough". An engine can only match those if the facts are written down somewhere it reads:

• Availability: emergency hours, same-day service, weekend calls, on your site and your Google Business Profile.
• Brands and equipment: the brands you service and install.
• Places: the cities and neighbourhoods you cover, by name.
• Specialties: heat pumps, boilers, backwater valves, tankless water heaters, whatever you actually do well.

4. Reviews get read, not just counted

When an AI explains why it suggests a company, it often summarizes reviews: "customers mention fast emergency service". Reviews that describe the job and the area are more useful than a star rating alone. Ask customers to say what you did and where. Don't script them, and don't reward them (Google's content policy bans that).

5. Follow-up questions favour clear pricing

AI conversations continue: "how much would that cost?", "should I repair or replace?" Businesses whose sites give honest price ranges and repair-versus-replace advice are more likely to be quoted in the follow-up, because they've published an answer.

What hasn't changed

• Your Google Business Profile still matters most for Google and Gemini. Our 10-minute checklist covers it.
• Consistency still matters. Same name, phone, website and service area everywhere. See our citations guide.
• Nobody can guarantee a spot. Answers vary by wording, place and day. Anyone promising "number one in ChatGPT" is guessing.
• Plenty of people still search the old way. The map pack and plain search results haven't gone anywhere. The good news is that the same work helps both.

Test it the way your customers ask

Spend 15 minutes on this before the next cold snap:

1. Write down five real questions from recent calls, in the caller's words, with the neighbourhood.
2. Ask each one in ChatGPT, Gemini and Google (in a private browser window).
3. Note who gets named, what's said about them, and which sources are linked.
4. Look for the gap: is it a missing page, a vague service area, missing hours, or fewer specific reviews?
5. Fix the biggest gap first, then ask the same questions again in a few weeks.

Keep your list of questions. Asking the same ones every season is the simplest way to see whether you're moving, and it shows you how your customers' questions are changing.

A note on seasons

GTA HVAC demand swings hard: heating questions spike with the first cold nights in October and November, and cooling questions with the first heat wave. Plumbing has its own peaks with spring rain and winter freezes. The questions homeowners ask AI follow the same pattern. Make your seasonal pages accurate a month before each peak, so they've been crawled by the time the calls start. Our Ontario HVAC calendar has the month-by-month plan.

For the specifics of each engine, see how to get recommended by ChatGPT and why AI Overviews skip some contractors.