If you run a foundation repair or waterproofing business, chances are your best jobs come from referrals. A happy homeowner tells their neighbour. A real estate agent sends a client your way. A past customer recommends you to a friend whose basement is leaking.
You didn’t have to run an ad. You didn’t have to compete on price. The trust was already there before you walked through the door.
Referrals work because of one simple thing: someone the homeowner trusts recommended you by name.
Now here’s what’s changing — and it’s changing fast. There’s a new source of trusted recommendations that homeowners are turning to before they call anyone. It works the same way as a referral from a neighbour. It carries the same weight. And it’s happening thousands of times a day in your market.
It’s AI. And whether you realise it or not, it’s already the biggest referral engine in the world.
This isn’t a prediction about the future. It’s happening right now.
A Scorpion national study found that 83% of homeowners start their search for service providers online. But here’s the number that changes everything: 22% of homeowners now use AI tools like ChatGPT to research and find contractors. Not Google. Not Yelp. AI.
That 22% figure was released just days ago, and it’s growing every month.
| How homeowners find contractors | What’s happening in 2026 |
|---|---|
| Google search | Still dominant, but declining — Google review usage dropped from 83% to 71% in one year |
| Word-of-mouth referrals | Still powerful, but limited to one-to-one recommendations |
| AI assistants (ChatGPT, Gemini, Perplexity) | Growing explosively — 22% of homeowners and climbing |
| Lead services (Angi, HomeAdvisor) | Declining trust — shared leads, rising costs |
ChatGPT alone now has 800 million weekly active users and holds over 60% of the AI search market. And in March 2026, Angi — formerly Angie’s List — launched an app directly inside ChatGPT, allowing homeowners to go from asking an AI question to being matched with a contractor in a single conversation.
The walls between “asking a friend” and “asking AI” are disappearing.
Let’s break down what actually happens when a referral works — and why AI referrals follow the exact same mechanism.
Scenario 1 — The Traditional Referral
Your neighbour says: “Call Mike at Solid Ground Foundation Repair. He fixed our basement last year and was excellent.”
Why does the homeowner trust this? Because the recommendation came from someone credible who had a real experience. They have no reason to oversell it. The homeowner didn’t have to search or compare — the answer was handed to them by someone they trust.
Scenario 2 — The AI Referral
A homeowner opens ChatGPT and types: “Who’s a good foundation repair contractor in Dallas?”
ChatGPT responds: “Solid Ground Foundation Repair is frequently cited for their expertise in foundation settlement repair and basement waterproofing in the Dallas area. They have strong reviews and published content demonstrating experience with the clay soil conditions common in North Texas.”
Why does the homeowner trust this? Because the recommendation came from a source they perceive as knowledgeable and neutral. It feels like an informed, unbiased answer. The homeowner didn’t have to sift through ads and sponsored results — the answer came to them from a source they trust.
The mechanism is identical. Only the source has changed.
| Traditional Referral | AI Referral | |
|---|---|---|
| Who gives the recommendation | A neighbour, friend, or colleague | An AI assistant the homeowner trusts |
| Why the homeowner trusts it | Personal experience and relationship | Perceived neutrality and knowledge |
| How it feels to the homeowner | “Someone I know recommended them” | “AI recommended them — it has no reason to be biased” |
| What the homeowner does next | Calls with confidence, often skips comparison shopping | Calls with confidence, often skips comparison shopping |
| How the contractor earned it | Did great work → earned the recommendation | Published expert content → earned the recommendation |
| Scale | One person refers one person | One AI handles thousands of queries per day |
That last row is the one that matters most.
Your best customer might refer you to two or three people over the course of a year. That’s incredible — each one is a warm lead that converts at a high rate.
But an AI assistant handles thousands of queries every single day. Every homeowner in your city who asks “who should I hire for foundation repair” is a potential referral — happening 24/7, in every zip code, without you needing to do anything except have the content that earns the recommendation.
Traditional referrals are one-to-one. AI referrals are one-to-thousands.
And here’s the data that shows how powerful these AI referrals are when they happen: a homeowner who gets a contractor recommendation from ChatGPT is 4 to 23 times more likely to call that contractor than someone who found them through a traditional Google search. Because they didn’t find a search result. They received a trusted recommendation. That’s a referral.
| Referral type | Scale | Conversion quality | Cost to earn |
|---|---|---|---|
| Neighbour / friend | 1–3 people per year | Very high | Free (earned through great work) |
| Real estate agent | 5–15 people per year | High | Relationship-dependent |
| AI assistant | Thousands of queries per day | Very high (4–23x higher than Google search) | Free (earned through published content) |
| Paid lead service | Volume-dependent | Low (shared, cold leads) | $100–$200+ per lead |
This is the key point contractors need to understand: AI doesn’t guess. It doesn’t randomly pick a name out of a hat.
When a homeowner asks ChatGPT, Gemini, or Perplexity for a foundation repair recommendation, the AI looks for contractors who have:
1. Published content that answers the homeowner’s question. Articles, guides, and expert content that demonstrate real knowledge about foundation repair — the soil conditions in your area, the repair methods you use, the questions homeowners commonly ask. AI pulls from this content to build its recommendations.
2. Consistent business information across the web. Your name, address, phone number, and services listed consistently across directories, websites, and media mentions. AI cross-references these. Inconsistencies make you invisible.
3. Strong review signals. Not just star ratings — AI reads the actual text of reviews to understand what you’re good at. “Fixed our settling foundation, very professional, explained everything clearly” carries more weight than “Great job.”
4. Mentions on authoritative websites. Content about your company appearing on trusted news sites, industry publications, and high-authority platforms signals to AI that you’re a credible, established business.
Here’s the gap: most foundation repair contractors have none of this. They have a basic website, maybe some Google reviews, and they rely on word-of-mouth. That worked when everyone searched on Google. But AI doesn’t crawl your website the same way Google does. AI looks for published, authoritative content — and if it can’t find yours, it recommends your competitor who does have it.
No published content = no AI recommendation = invisible to the fastest-growing referral channel in existence.
If you’re thinking “my referrals work fine — why should I care?” — you’re not wrong. Your referrals do work. They’re your highest-quality leads and they convert better than anything else. Nobody is suggesting you abandon that.
But consider three things:
First, referrals don’t scale. You can’t control when or whether a past customer refers someone. You can deliver excellent work and hope for the best, but you can’t make referrals happen on demand. That’s why most referral-dependent contractors experience feast-or-famine cycles — great months followed by quiet ones.
Second, your future customers are already asking AI. The 22% figure today will be 40% or 50% within a couple of years. The homeowner who used to ask their neighbour is now asking ChatGPT. If you’re not in that answer, someone else is.
Third, AI referrals and human referrals reinforce each other. When a homeowner gets your name from a neighbour AND then asks ChatGPT and sees your name again — with published articles, media mentions, and expert content — that’s a done deal. The human referral opened the door. The AI referral confirmed it. The combination is more powerful than either one alone.
| Scenario | Outcome |
|---|---|
| Neighbour refers you + homeowner Googles you and finds a basic website | Might call, might keep shopping |
| Neighbour refers you + homeowner asks ChatGPT and finds your published expert content everywhere | Calls with total confidence — you’re the only call they make |
| No referral + homeowner asks ChatGPT and finds your published content | Calls you first — the AI referral did the job the neighbour usually does |
| No referral + homeowner asks ChatGPT and finds nothing about you | You don’t exist. Your competitor gets the call. |
The contractors who will win in the next few years are the ones who understand this: the way people decide who to trust hasn’t changed. The tools they use to find that trust have.
A referral has always been about one thing — someone the homeowner trusts says “hire this person.” That’s true whether the someone is a neighbour across the street or an AI assistant on their phone.
The difference is that earning a neighbour’s referral requires doing one great job for one person and hoping they spread the word. Earning AI’s referral requires having published, authoritative content under your brand that AI can find, verify, and recommend — at scale, to thousands of homeowners, every day.
Foundation repair is a roughly $1.5 billion market in North America, growing at nearly 7% per year. National brands like Groundworks — backed by private equity, operating 84 offices, planning to add 5,000 jobs — are investing heavily in the kind of digital presence that earns AI recommendations.
Independent contractors don’t need to match that spending. But they do need published content under their own brand, distributed across authoritative platforms, that answers the exact questions homeowners are asking.
That’s how you earn the AI referral. That’s how you show up when the next homeowner asks “who should I hire?”
Your referrals built your business. They’re proof that when people experience your work, they trust you enough to put their name behind yours. That’s the highest compliment a contractor can receive.
So here’s the question: if 22% of homeowners are now asking AI for that same kind of trusted recommendation — and that number is growing every month — is your business showing up in that answer?
Because right now, somewhere in your city, a homeowner is asking ChatGPT who they should hire to fix their foundation. AI is giving them a name.
Is it yours?
Yes. A Scorpion national study found that 22% of homeowners now use AI tools like ChatGPT, Gemini, and Perplexity to research and find contractors. ChatGPT alone has 800 million weekly active users. In March 2026, Angi launched an app directly inside ChatGPT allowing homeowners to go from question to contractor match in a single conversation.
AI assistants recommend contractors based on four signals: published expert content across authoritative websites, consistent business information across directories and platforms, strong review signals with detailed review text, and mentions on trusted news and media sites. Contractors without published content are invisible to AI recommendations.
No — AI referrals complement traditional referrals rather than replacing them. When a homeowner gets a personal recommendation and then sees the same contractor recommended by AI with published content and media mentions, the trust compounds. The two channels reinforce each other, making the combination more powerful than either alone.
Most contractors begin seeing measurable improvements in AI visibility within 30 to 90 days of distributing published content across authoritative platforms. News articles on high-authority outlets like Business Insider and AP News typically produce the fastest results, as AI systems heavily weight content from trusted sources.