ChatGPT for Plumbers

ChatGPT for Plumbers: The 90-Minute Window You're Losing to Your Contact Form

When a homeowner stares at a flooded kitchen at 2am, they don't open Google — they open ChatGPT. Whichever plumber is reachable in the chat wins the call. Here's how that flow actually works, and the math on what it's worth.

Updated · Sav Pushparajah

The pitch in one paragraph. Homeowners with a leaking ceiling are typing into ChatGPT, not Google. ChatGPT recommends two or three local plumbers by name. The next move is supposed to be: leave the chat, find the website, fill out an emergency contact form. They don’t. They call the first plumber whose number is plainly visible — usually not yours. With MyDeetz, the homeowner shares their address, phone, and the problem inside the conversation, and the lead is on your dispatcher’s screen in under 30 seconds. The form was the leak. This is the play.


The 2am moment that decides who gets the call

It’s 2:14am. There’s water coming up through a kitchen tile. The homeowner is barefoot, panicking, and they’ve already turned the wrong valve. They unlock their phone and open the app they trust to give them a fast answer to anything: ChatGPT.

They type: “there is water coming up through my kitchen floor what do I do.”

ChatGPT does two things. First, it walks them through shutting off the main water valve (which calms them down enough to keep typing). Then it says: “You’ll want a plumber out tonight. In your area, plumbers with strong reviews and 24/7 dispatch include…” — and it names two or three.

If you’re one of the names, you have about 90 minutes to be the one they actually call. That window closes when (a) someone else picks up the phone, (b) they fall asleep, or (c) they get distracted and decide to “deal with it in the morning” — and by morning the urgency, and the willingness to pay an emergency rate, are both gone.

This page is about owning that window. It’s not another “do better SEO” post. The leverage isn’t being named by ChatGPT. It’s being contactable in the conversation where the naming happens.

Why plumbing is the highest-leverage trade for this

Three things make plumbing different from, say, landscaping or general handyman work:

  1. Emergency premium is real. A 2am burst-pipe call is $400–$1,200 of work, often more once they see the damage. A scheduled tankless install is $4,000–$8,000. The unit economics on AI-sourced inbound are excellent because the people using AI in panic mode are high-intent, time-pressed, and willing to pay.

  2. Decision window is minutes, not days. Nobody shops 14 plumbers when their basement is flooding. They pick the first responsive one. Speed of capture is the entire game.

  3. Local Services Ads cap your reach. LSAs work, but you’re paying $25–$150 per qualified call, capped to your daily budget, competing on a shrinking list of slots, and you don’t get the lead at all if your phone rings while you’re under a sink. ChatGPT capture is additive — it grabs the leads that would have otherwise gone to a competitor or evaporated.

This is also why high-urgency trades — emergency plumbing, HVAC after-hours, electrical, water-damage restoration — are where AI search is replacing Google search fastest. People with urgency talk to a chatbot. People doing research read listings.

What the in-chat flow actually looks like

With MyDeetz installed, the literal flow is:

  1. Homeowner: “There’s water coming up through my kitchen floor, I need a plumber tonight in [city].”
  2. ChatGPT recommends 2–3 firms in their service area, including yours.
  3. Homeowner: “Send my details to [Your Plumbing Company].” (or @mydeetz send my details to ...)
  4. ChatGPT collects the fields you configured — name, phone, address, problem description, urgency, optional photo description.
  5. The lead lands in your dispatcher’s email and triggers a webhook to your on-call tech’s phone via your CRM (ServiceTitan / Jobber / Housecall Pro), within 30 seconds.
  6. Your tech calls them back inside 5 minutes, while the water is still on the floor and the credit card is still on the kitchen counter.

No website visit. No emergency contact form. No “we’ll get back to you within 1 business day.” No tab switch back to Google.

What the form is actually costing you

Take a small residential plumbing company doing $1.2M/year in revenue, with about 35–40% from emergency calls. Conservative assumptions:

  • 400 emergency-class inbound prospects per month from all channels.
  • 80 of those are arriving via ChatGPT or other AI search by mid-2026 (and rising fast — homeowners 25–45 default to ChatGPT for after-hours panic queries).
  • Form-completion rate on the firm’s emergency contact page: 28% (low because forms in panic mode lose to phone numbers).
  • Of completed forms that actually pick up the phone when called back: 40%.
  • Of calls picked up that book a job: 55%.

That’s 80 × 0.28 × 0.40 × 0.55 = ~5 booked emergency jobs/month from AI-sourced inbound. At an average emergency ticket of $620, ~$3,100/mo, ~$37k/year.

Now replace the form with in-chat capture. Realistic assumptions from our pilot home-service contractors:

  • “Form-equivalent” completion in chat: 75% (no fields to type, no page to load, the homeowner already typed their problem to ChatGPT).
  • Pickup rate on callback: 65% (the homeowner is sitting next to the phone they just used to capture).
  • Booking rate: 60% (slightly higher because they pre-described the problem so the call is shorter and more decisive).

80 × 0.75 × 0.65 × 0.60 = ~23 booked emergency jobs/month. At $620, ~$14,300/mo, ~$172k/year.

That’s an additional ~$135k/year for the same top-of-funnel volume. The variable wasn’t more leads. It was less leak.

(If you do commercial work, the numbers are more lopsided — fewer leads but each one is bigger. A single backed-up restaurant kitchen on a Friday night that you respond to in 4 minutes instead of 4 hours is a $3,000–$8,000 ticket you would have lost to whoever the property manager texted first.)

What plumbing companies actually need to do

In priority order. Items 1–3 are AEO foundations every trade needs. Item 4 is what this page is about.

1. Make sure ChatGPT can find and read your site. Your robots.txt should allow GPTBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended, and Applebot-Extended. Most plumbing sites built by old SEO agencies still default to blocking these. Audit it once, fix it in 5 minutes, never think about it again.

2. Get the right schema on the page. Schema.org has a dedicated Plumber type — yes, really. Use it on your homepage and service pages. Add LocalBusiness with openingHours (mark emergency hours separately as validFrom: 00:00, validThrough: 23:59). Add FAQPage schema on each major service page — this is what AI engines mine for direct-answer extraction. Add Person schema on your owner/lead-tech bio with sameAs pointing to your Google Business Profile, Master Plumber license lookup, and any trade association membership.

3. Dominate Google Business Profile + reviews. GBP is still the highest-weight third-party signal AI engines use to validate trades. Concrete:

  • Primary category: Plumber. Secondary: Emergency Plumbing Service, Drain Cleaning Service, Water Heater Store, Sewer Repair.
  • Service-area set per ZIP, not just “city” — granularity matters for local-intent AI queries.
  • Hours: emergency availability marked clearly, with the after-hours number live.
  • 100+ Google reviews at 4.7+ average. Texting customers 2 hours after job completion via ServiceTitan/Jobber automations gets 30–50% conversion rate.
  • Weekly posts during seasonal events (“frozen pipe checklist,” “summer water-heater inspection deal”).

4. Be reachable inside the conversation. Register your company on MyDeetz at app.mydeetz.ai. Pick the fields your dispatcher actually needs to triage a call: name, phone, address, problem description, urgency (timeline), and optionally a photo description. Set the delivery email to your dispatcher’s inbox, and on the Business plan, fire a webhook into your CRM so an emergency tag instantly pages your on-call tech. Add the @MyDeetz deep link to your site footer and your GBP description.

The first three buy you long-term placement in the candidate list. The fourth converts the prospect while they’re still standing in the puddle.

”But Local Services Ads work for us”

Good. Keep them. MyDeetz isn’t replacing LSAs — it’s replacing the contact form on your website.

The leads that come from LSAs are leads that picked you because Google placed your phone number in front of them. Those are still yours. The leads from ChatGPT are leads who never saw your LSA — they bypassed Google entirely. Today, those leads either dial whichever plumber’s name was easier to act on, or they evaporate.

LSA spend goes up over time as competition increases. AI-sourced inbound is in a 12–24 month early-mover window where almost no competitor has installed a capture path. The unit economics on the AI-sourced lead are dramatically better.

”We use ServiceTitan / Jobber / Housecall Pro”

Good. MyDeetz isn’t a CRM — it’s the channel that feeds the CRM. Today, your CRM gets fed by your website’s contact form, phone calls, and LSA leads. Tomorrow, it’s also fed by leads from ChatGPT. Same downstream workflow, new top-of-funnel source.

On the Business plan ($149/mo), urgent leads (where the homeowner indicated timeline = right now or today) fire a webhook with a custom payload, which any of those CRMs can ingest to auto-page your on-call tech via SMS. We’re working on first-class ServiceTitan and Jobber integrations — until then, the webhook plus a Zapier hop covers it in about 15 minutes of one-time setup.

The early-mover argument

Most plumbing companies in 2026 are still doing two things:

  1. Spending more on Local Service Ads as the per-call cost climbs.
  2. Talking about getting around to “redesigning the website” sometime this year.

Almost nobody is thinking about the contact-step in the AI-search flow yet. That window — between “AI search is real” and “every competitor has installed an AI lead-capture tool” — is roughly 12 to 24 months. Right now, you can be one of the few plumbing companies in your metro that doesn’t lose the lead at the form. By 2027, this is going to be table-stakes the same way “answer the phone within three rings” became table-stakes once everyone had a smartphone.

The unfair advantage isn’t being smart. It’s being early.

FAQs

Does MyDeetz actually route emergency calls fast enough? Lead delivery is sub-30 seconds (email + webhook on the Business plan). Pair the webhook with a Zapier or ServiceTitan/Jobber automation that texts your on-call tech the moment a timeline = urgent lead arrives. Most pilot contractors are calling the homeowner back inside 4 minutes.

Do I still need Google Local Service Ads? Yes. LSAs are still excellent for plumbers. They’re a different channel — they capture intent on Google. MyDeetz captures intent on ChatGPT. Both at once is the right move; one doesn’t replace the other.

What about HomeAdvisor / Angi? Mixed. Many plumbers are reducing or cancelling those subscriptions because per-lead costs keep rising and lead quality keeps dropping (shared leads, tire-kickers). MyDeetz costs $49–$149/mo flat for unlimited leads, which is structurally cheaper if AI-sourced volume keeps growing.

Do ChatGPT-sourced leads close at a higher rate? In our pilot data with home-service contractors, AI-sourced leads close at roughly 1.5× to 1.8× the rate of generic web-form leads. The homeowner has already described the problem to ChatGPT, so by the time you call them back, they’re warmed up, they trust that you know what you’re walking into, and the call is shorter. They also tend to be in the urgent-decision state where “schedule it now” is the path of least resistance.

What if I only do commercial plumbing? Same playbook, different intake fields. Add company, role, and company_size to the capture so a facilities manager at a commercial property gives you everything a dispatch needs in the first message. Commercial leads tend to be higher-value and more deliberate, but the same “in-chat capture beats a contact form” thesis applies — facilities managers are using ChatGPT more than they’re using your website.

How long until I see results? Schema and robots.txt fixes: AI engines re-ingest within 1–4 weeks. Google Business Profile improvements: 2–6 weeks for AI engines to weight them. The MyDeetz contact-step fix works immediately because it doesn’t depend on ranking — the moment a homeowner chooses to share their details with you in chat, the lead arrives.

Is MyDeetz the same as a chatbot on my website? No. A website chatbot only works if the prospect arrives at your website. MyDeetz works before they arrive — it intercepts the prospect in ChatGPT, where they were going to make their decision. Most AI-sourced prospects never reach your website at all.


Get reachable in 10 minutes

  1. Sign up at app.mydeetz.ai. Free plan, no card.
  2. Pick your company name (the slug ChatGPT will use to recognise you) and your dispatch fields — name, phone, address, problem, urgency.
  3. Set delivery to your dispatcher’s email, plus on Business, a webhook into your CRM for after-hours auto-paging.
  4. Add the @MyDeetz deep link to your website footer and your GBP description. (Optional but it accelerates AI citation by 2–4 weeks.)

That’s it. The next time a homeowner in your service area asks ChatGPT for an emergency plumber, they can choose to share their details with you inside the chat. The lead lands in your dispatcher’s hand within seconds.

PlanCostWhat you get
Starter$0/mo2 leads/month, default fields (name/phone/email), 1 delivery email
Pro$49/mo (or $33/mo annually)Unlimited leads, full 13-field custom catalogue (incl. address, urgency, problem description), 3 delivery emails, AI Discoverability page (your /business/<slug> becomes a real, AI-optimised landing page)
Business$149/moUnlimited leads, 10 emails + webhook URL (the right plan for emergency dispatch routing), priority onboarding, CRM integrations coming (ServiceTitan + Jobber on the roadmap)

The number to beat isn’t your form-completion rate. It’s the number of homeowners who never see your form at all because they made their decision inside ChatGPT. That’s the leak this fixes.


Bottom line

ChatGPT will name your plumbing company. The AEO basics (schema, robots, GBP, reviews) are how that happens — and you should keep investing in them. But the naming isn’t the moment that converts. The contact-step is. Today, the contact-step is a website visit and an emergency contact form, and the majority of high-intent homeowners don’t survive it at 2am.

Be reachable in the conversation. The plumbing companies in your metro who figure this out first will absorb the AI-sourced inbound that everyone else is leaking.

Setup takes about 10 minutes. The leak you close runs forever.

Get leads from ChatGPT — free