HVAC and Plumbing Marketing in Zurich 2026: Why ChatGPT, Gemini and Google AI Mode Recommend Your Building Services Firm, or Don't

As of October 2026. Measured on 13 October 2026.

For heating, ventilation, air conditioning and plumbing (HLKS in Swiss German), the field is more open than in the other industries in our October measurement. We asked ChatGPT, Claude, Gemini and Google AI Mode eight typical questions about HLKS firms in Zurich: 32 answers, 173 mentions, 102 different firms. The five most-named get only 17 percent of mentions, and 69 firms appeared exactly once. What gets cited is mainly the firms' own websites, especially pages combining a service and a place, emergency pages and subsidy pages. And there is a gap: all eight answers about heat pumps talk about subsidies, not one about the Zurich energy law.

"Wärmepumpe" (heat pump) is searched on Google around 6,600 times a month in Switzerland, "Sanitär" (plumbing) 5,400 times, "Heizung" (heating) 3,600 times, "Sanitär Zürich" 1,000 times. Emergencies are the most expensive: Google Ads estimates around USD 22 per ad click for "Sanitär Notfall Zürich" and around USD 11 for "Sanitär Zürich". Then there are the big jobs: according to the AI answers themselves, replacing a heating system with a heat pump quickly costs CHF 30,000 to 60,000. Anyone who has to replace an oil heating system today often asks an AI first which firm to consider.

For this article, on 13 October 2026 we put eight questions to four assistants, ChatGPT, Claude and Gemini with web search and location Zurich, Google AI Mode with location canton of Zurich:

• "Welche Firma in Zürich kannst du für den Ersatz meiner Ölheizung durch eine Wärmepumpe empfehlen?"
• "Heizungsinstallateur in Zürich: Wer ist empfehlenswert?"
• "Sanitär Notfall Zürich: Welcher Sanitär ist schnell und zuverlässig?"
• "beste haustechnik firma zürich"
• "Ich will in meinem Einfamilienhaus im Kanton Zürich eine Wärmepumpe einbauen. Welche Anbieter kommen infrage und was kostet das ungefähr?"
• "Nenne mir 5 gute HLKS-Firmen in Zürich und was für sie spricht."
• "Klimaanlage oder Komfortlüftung in Zürich einbauen: Welche Firma ist gut?"
• "Sanitär und Heizung Winterthur Empfehlung"

Who gets named: a hundred firms, nobody dominates

Claude named around seven firms per answer on average, Gemini six to seven, Google AI Mode four to five, ChatGPT three. Of the 173 mentions, 119 went to local installers, 24 to larger building services groups, 13 to emergency services, 10 to heat pump specialists and 5 to energy utilities.

The five most-named account for only 17 percent of all mentions. By comparison: for bathroom and kitchen renovation it was 21 percent, for aesthetic medicine 22, for real estate agents 44 percent. Only two firms were named by all four assistants.

What is striking is how much the names depend on the question. For the Winterthur question, all four assistants named Lyrenmann; for the emergency question, three of four named HT Haustechnik. For the question about replacing an oil heating system, ecoEn, a provider specialising in heat pumps, appeared in three of four answers, each time with its heat pump page as the source. For the heat pump cost question, the assistants hardly named installers at all, but rather specialists and energy utilities such as EKZ or Energie 360°, and Google AI Mode named no firm at all. For a local business this means: there is no single ranking. Whoever answers a question best gets named for exactly that question.

What the AI wants to hear: availability, advice, proximity

The assistants' most frequent reasons were reviews, planning and advice, regional proximity, emergency service and availability, maintenance and service, heat pump experience and "everything from one source". Price hardly played a role as a reason.

Many answers also give homeowners advice: 21 of 32 talk about quotes, and 13 explicitly recommend getting several. 16 mention a 24-hour emergency or on-call service, 12 maintenance or service contracts, 7 the industry association suissetec. The heat pump quality label WP-System-Modul came up in 4 answers.

For a firm's website this means: the terms the assistants use to justify their picks should be there as concrete statements. "24-hour on-call service for Zurich and the Glattal, usually on site within 60 minutes" can be quoted; "your reliable building services partner" cannot.

Where the AI gets its recommendations

62 percent of the cited pages are firm websites, 20 percent portals and Google Maps, 11 percent authorities, energy utilities and associations, 6 percent lists and rankings. The share of authorities, utilities and associations is unusually high: for real estate agents, bathroom renovators and clinics it was between 3 and 6 percent. The reason is the heat pump: for questions about replacing a heating system, the assistants read the canton's subsidy programme and the pages of the City of Zurich and the energy utilities.

Among the firms, it is mainly pages about a specific service that get read. Of 117 cited company pages, 47 had a service in the address, such as heating, plumbing, emergency, heat pump or air conditioning, and 27 a place, such as "heizung-zuerich", "sanitaer-notfall-zuerich" or "klimaanlagen-in-zuerich". 50 were home pages. Among the most-cited company pages were an emergency plumbing page and a heat pump page, each in 3 answers.

The subsidy pages stand out. Five firm pages on "subsidies for heat pumps in the canton of Zurich" were cited, plus the canton's subsidy brochures. Whoever answers the subsidy question on their website ends up as a source in the answer, often with their name.

Among portals, daibau.ch, local.ch, search.ch, ofri.ch and handwerker.ch were cited most often. A complete profile there does no harm, but the bigger lever is your own website.

Emergency service: AI checks who is really local

The question about a plumbing emergency shows how closely the assistants look. In its answer, Claude explicitly distinguished between firms with their own base in the region and pure emergency hotlines where it was not clear whether and where a permanent team works. Its tips for users: ask about the hourly rate, call-out fee and night surcharges in advance, check where the firm is based, read the Google reviews. ChatGPT also mentioned a fixed-price emergency offer from the TCS.

For a firm's emergency page this means: state your base, your own team, service area, response time, hourly rate, call-out fee and surcharges openly. These are exactly the details the assistants look for, and they separate a genuine local firm from a referral hotline.

Heat pumps: costs, subsidies and a gap

For the cost question, the assistants give concrete ranges: for an air-to-water heat pump in a detached house mostly CHF 25,000 to 48,000, for a ground-source heat pump CHF 45,000 to 80,000 and more, of which around CHF 15,000 to 25,000 is for drilling. Some answers list removing the oil heating, disposing of the tank and adapting the radiators separately. Google AI Mode, incidentally, wrote its amounts with dollar signs, even though they were Swiss francs.

All eight answers to the two heat pump questions mention subsidies. None of the 32 answers, however, mentions the Zurich energy law. Yet it is central for homeowners: since 1 September 2022, fossil heating in existing buildings in the canton of Zurich may in principle only be replaced with renewable systems. Exceptions apply if this is technically impossible or would be more than 5 percent more expensive over the life cycle. According to a review by the cantonal building department in ten municipalities, more than 98 percent of the fossil heating systems replaced in the first year were replaced with climate-neutral systems, mostly heat pumps.

That is an opportunity for every heating firm. A page "Replacing an oil heating system in the canton of Zurich: what the energy law requires, what it costs and which subsidies are available" answers exactly the questions owners ask, and gives the assistants content they currently find nowhere. Important: subsidy rates change, so add a date and a link to the canton's subsidy programme.

Google AI Mode and ChatGPT Ads

Google AI Mode relied on Google Maps data for 6 of 8 questions. The suggested firms had a median of 4.8 stars and around 20 reviews, one of them just a single review. So you do not need many reviews, but you do need a complete profile with the right categories ("Plumber", "Heating contractor", "Air conditioning contractor"), service area, opening hours and emergency service. How Google AI Mode selects across all industries is shown by our measurement with 100 questions.

Building services firms may advertise in ChatGPT: local services are one of the permitted categories, and ads can be targeted to cities and postcodes, from CHF 20 per day. Especially for the expensive emergency search term, that is an alternative. How it works is covered in ChatGPT Ads in Switzerland.

How do I get ChatGPT and Google to recommend my HVAC or plumbing business?

1. One page per service and place. "Heizung Zürich", "Sanitär Winterthur", "Wärmepumpe Zürcher Oberland", "Klimaanlage Zürich". With place names in the title, in the text and in the references.
2. An honest emergency page. Base, own team, service area, response time, hourly rate, call-out fee and surcharges. That shows you are a genuine local firm.
3. A page on heating replacement. Energy law, process, cost ranges for air-to-water and ground source, subsidies with date and link. This is the gap no answer fills today.
4. Maintenance and service as a service in its own right. Service plan, intervals, prices. Maintenance came up in 12 answers.
5. References with location and scope. "Oil heating replaced with an air-to-water heat pump, detached house in Uster, 12 kW, 3 days" tells the assistant more than a strip of logos.
6. Quality signals as text. suissetec membership, WP-System-Modul, manufacturer partnerships, training company. As text, not just as a logo in the footer.
7. A complete Google Business Profile. Categories, service area, emergency hours, project photos, reviews after every job.

How recommendations come about in general is described in How do I get recommended by ChatGPT?. What this looks like in practice for an energy provider is shown in our case study with a solar provider.

Templates you can copy

The following templates put the data into practice. Replace the placeholders in square brackets with your own details, and only write what is true for your firm.

Page "Replacing an oil heating system in the canton of Zurich"

Page title:  Ölheizung ersetzen im Kanton Zürich: Pflicht, Kosten, Förderung | [Firm]
URL:         /oelheizung-ersetzen-zuerich

H1   Replacing an oil heating system in the canton of Zurich
     Three-sentence intro: who the page is for, what it answers,
     who you are and which municipalities you cover. As of: [month year].
H2   What the Zurich energy law requires          (box, see below)
H2   Which systems are an option                  air-to-water, ground source, district heating
H2   What the replacement costs                   ranges from your own projects
H2   Which subsidies are available                table, see below
H2   Process and duration                         site visit, quote, application,
                                                  installation, commissioning
H2   References from the region                   place, building, system, kW, duration
H2   Frequently asked questions                   5 to 8 real customer questions

Write the page in German for your Zurich customers; the English outline is only here for readability.

The energy law box, factual and with a source:

The subsidy table, with a date and without amounts that go out of date:

Emergency page: the details that belong on it

Structured data for an HVAC and plumbing firm

This JSON-LD template goes in the <head> of the home page and the emergency page. The details must match the visible text.

{
  "@context": "https://schema.org",
  "@type": ["Plumber", "HVACBusiness"],
  "name": "[Firm]",
  "url": "https://[your-domain].ch",
  "telephone": "+41 [number]",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "[street no.]",
    "postalCode": "[postcode]",
    "addressLocality": "[town]",
    "addressCountry": "CH"
  },
  "areaServed": ["Zürich", "[municipality]", "[municipality]"],
  "openingHoursSpecification": [{
    "@type": "OpeningHoursSpecification",
    "dayOfWeek": ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday", "Sunday"],
    "opens": "00:00",
    "closes": "23:59"
  }],
  "memberOf": { "@type": "Organization", "name": "suissetec" },
  "sameAs": ["[Google Business Profile link]", "[LinkedIn link]"]
}

Only include round-the-clock opening hours if your emergency service really is reachable around the clock, and the membership only if you are a member.

Checklist: 10 points for your HLKS firm's website

And an eleventh point: measure it. Ask ChatGPT, Gemini and Google AI Mode the same five questions every month, with your region and your services, and note who gets named. How to do that is explained in measuring AI visibility.

What we do for building services firms

Hierarchy is a full-funnel marketing and AI agency for SMEs in Zurich, specialised in SEO, GEO and paid ads, with marketing and engineering in a team of four. For HLKS firms, we build service and location pages, emergency and heating replacement pages so that Google and the AI assistants find and cite them, maintain the Google Business Profile and, if you wish, run Google Ads and ChatGPT Ads for your region. Every month you see whether your firm is named. Where you stand today is shown by the free AI Visibility Check. Details: SEO & AI visibility and paid ads.

Conclusion

For heating, ventilation, air conditioning and plumbing, AI recommendation is more open than in the other industries in our October measurement: 102 firms in 32 answers, and the five most-named get only 17 percent. Which names come up depends heavily on the question: the place, an emergency, a heat pump. What gets read is mainly the firms' own websites, especially pages combining a service and a place, emergency pages and subsidy pages. Whoever writes these pages honestly and concretely, and is first to explain the energy law clearly, supplies exactly the material the assistants build their recommendations from. How other industries fare: bathroom and kitchen renovation, architecture firms and real estate agents.