Zero click search now means a customer asks ChatGPT, Gemini or Google AI Overviews for a recommendation and gets a shortlist of named businesses without visiting a website. The lead is won or lost inside the answer. To capture it, you must be the business the AI names, backed by facts it can verify.
Marketers used to mean something smaller by "zero-click." A featured snippet answered a question and you lost a visit. Today the question is different. A homeowner asks an assistant who to call about a leaking water heater in their area. The assistant replies with two or three business names, a reason for each, and often a phone number. That customer was never going to read your website. They were going to pick from a shortlist, and the shortlist was assembled before they arrived.
TruLata is an AI marketing platform for service businesses, including home services companies, law firms, clinics, dental practices and advisory firms. It measures whether ChatGPT, Gemini, Claude, Grok, Manus, Google AI Overviews, Perplexity, Meta AI and DeepSeek name a business when customers ask. This guide sets out the strategy for zero click search now that the answer, not the website, is where the shortlist forms.
What did zero click search used to mean, and what does it mean now?
The old version: lost clicks to Google features
The first wave of zero click search came from Google itself. Featured snippets, knowledge panels, "People also ask" boxes and the local pack answered simple questions on the results page. Someone searching for a business's opening hours saw them and left. The cost was mostly informational traffic: visits that rarely turned into revenue.
The standard response was to win the snippet. You structured content so Google lifted your answer, accepted fewer clicks, and treated the brand impression as a consolation prize.
The new version: lost leads to AI recommendations
The second wave is a different kind of problem. AI assistants do not just answer facts. They make recommendations. When someone asks "who is the best emergency plumber near me" or "which roofer should I call after hail damage," the engine compares options and names a few businesses.
The question is a hiring decision, not research. If your business is not named, you are not in the running. If it is named, the customer may call you directly without ever loading your homepage. In both cases your website analytics show nothing, because nothing happened on your website.
That is why the strategy has to change. Winning a snippet was about earning visibility for a page. Winning an AI recommendation is about being the business the engine trusts enough to name.
How many searches now end without a click?
The shift is no longer theoretical. Research from Bain & Company found that about 80% of search users rely on AI summaries or other zero-click results in at least 40% of their searches. Bain estimates this reduces organic web traffic by 15% to 25%. The finding comes from the Bain and Dynata Generative AI Consumer Survey of 1,117 respondents in December 2024.
For local service providers the effect is sharper, because their customers have been trained to expect options on the results page. Search Engine Land's guide to zero-click searches observes that for many local businesses, especially service providers, zero-click visibility has replaced traditional website traffic as their primary digital acquisition channel. People increasingly search for the service plus "near me" rather than a business name, and expect the engine to deliver credible options.
The practical consequence is simple. A drop in website sessions does not have to mean a drop in demand. It may mean demand is being decided somewhere you are not measuring.
Why do AI recommendations for home services skip your website?
The customer wants a name, not an article
AI recommendations for home services serve a customer who has a problem now: a dead furnace, a burst pipe, a missing shingle. They do not want to compare ten websites. They want two or three trustworthy names and a reason to choose one. The assistant gives them exactly that, and the conversation often ends with a phone call.
That is good news, in one sense. The action the customer still needs to take (calling, booking, getting a quote) cannot happen inside the answer. The click that remains is high intent. You just have to be one of the names attached to it.
The engine builds the answer from what the web says about you
Answer engines assemble recommendations from signals spread across the web: your Google Business Profile, your reviews, your listings in directories, your own service pages and mentions elsewhere. Search Engine Land notes that the businesses winning local visibility do it through optimized Google Business Profiles, consistent citations and strong review signals. Those same inputs feed the AI layer.
This is the same pattern behind roofing leads disappearing into answer boxes, and it applies to every trade where customers hire locally. Your website still matters, but as a source the engine reads, not as a destination the customer visits.
What do AI engines need before they will name your business?
No engine publishes a formula for who gets recommended, so be wary of anyone who claims to know one. What you can control are the inputs that make a business easy to understand, easy to verify and low risk to recommend.
Identical facts everywhere
Your business name, address, phone number, hours, service area and list of services should match across your website, your Google Business Profile and every directory listing. Conflicting details give an engine a reason to hesitate. Audit the top listings for your trade and region, fix what is out of date, and keep a single master record of the facts.
Reviews that describe real jobs
A rating tells an engine you are liked. Review text tells it what you are liked for. A review that mentions "replaced our tankless water heater in two hours" gives the engine something specific to match against a question about water heaters. Ask every satisfied customer for a review, and make it easy for them to mention the job and the neighbourhood. Respond to reviews, including the critical ones, with calm and specific replies.
One clear page per service and area
A single "Services" page listing twelve trades is hard to cite. A dedicated page for each core service, written in plain language, answers the questions an engine is trying to resolve: what you do, where you do it, how a job typically runs, and how to reach you. Write the first paragraph of each page as a direct answer that makes sense on its own, because that is the span most likely to be lifted.
Claims you can source
Engines increasingly cross-check what a business says about itself. An award you cannot point to, a review count that does not match your profile, or a "25 years in business" line that contradicts your registration date all erode trust. TruLata's website builder is built around this risk: its patent-pending claim-verification technology will not publish a rating, review count, year in business, credential or award without a source on file. Words the owner types become a dated source, and claims the AI invents are refused.
How do you capture zero-click conversions when nobody visits your site?
Zero-click conversions are leads that come from an AI answer or results page without a site visit in between. They arrive as phone calls, profile messages, direction requests and bookings. Capturing them takes a few deliberate changes.
- Make contact the primary action everywhere you appear. Your phone number, booking link and hours should be accurate and prominent on your Google Business Profile and every listing. If an engine surfaces your number, it should ring a person or a reliable answering service.
- Answer fast. A customer who got your name from an assistant also got two competitors' names. If you miss the call, the next name on the list is one tap away.
- Ask every new lead how they found you. Add "an AI assistant such as ChatGPT or Gemini" as an option on your forms and in your call script. Self-reported source is imperfect, but it is often the only signal a zero-click lead leaves.
- Count leads where they actually land. Platform reports and analytics tools often miss leads that skip the website. TruLata counts leads from your own forms, server-side, with each lead's source and an outcome your team sets.
- Watch the indirect signals. Branded searches, profile calls and direction requests tend to rise when AI engines start naming you. Track your rankings, queries and reviews side by side so a change in one can be read against the others.
How should you measure visibility in AI answer boxes?
Traffic was the old proxy for visibility. In AI answer boxes, the measure is whether you are named, how often, and for which questions. Search Engine Land frames this as the central measurement problem of zero-click search: success no longer shows up in the metrics teams were built to report.
Build a fixed prompt set
Write down 15 to 30 questions your customers actually ask, in their words. Mix emergency queries ("who can fix a leaking roof today"), comparison queries ("best HVAC company in my city for older homes") and service queries ("who installs heat pumps near me"). Keep the list stable so results are comparable over time.
Check every engine, not just one
Customers do not all use the same assistant. One engine may name you consistently while another never does, and the fix differs depending on which sources that engine leans on. Running the same prompts by hand across several engines every week is slow, which is why TruLata measures whether nine answer engines name you: ChatGPT, Gemini, Claude, Grok, Manus, Google AI Overviews, Perplexity, Meta AI and DeepSeek.
Put visibility next to leads
Being named is a leading indicator. Leads are the outcome. Review both on the same screen and look for patterns. Did profile calls rise in the weeks after you started appearing in Gemini answers? Did self-reported "AI assistant" leads climb after you published dedicated service pages? You will rarely get a clean one-to-one attribution, but a consistent pattern across signals is a credible measure.
What does a 30-day zero click search plan look like?
You do not need a large project to start. A focused month covers the foundations.
- Week 1: Baseline. Write your prompt set. Run it across the major engines and record who is named, including competitors. Note the reasons each engine gives for its picks.
- Week 2: Fix the facts. Correct your name, address, phone, hours and service list everywhere they appear. Remove or source any claim on your website you cannot back up.
- Week 3: Strengthen the sources. Publish or rewrite one page for each core service, each opening with a direct answer. Start a steady review request routine that asks customers to mention the job.
- Week 4: Wire up measurement. Add the "AI assistant" source option to forms and call scripts. Make sure every form submission and profile call is counted. Rerun the prompt set and compare.
After the first month, repeat the prompt set on a schedule, and treat movement in AI answers as a weekly metric alongside calls and booked jobs.
How does TruLata help you get named and count the leads?
Zero click search rewards businesses that can see the whole picture: whether the engines name them, what customers say in reviews, how they rank, and which enquiries actually arrive. TruLata puts those on one screen that refreshes when you open it. It tracks your visibility across nine answer engines, shows your search rankings and reviews, counts leads from your own forms, and runs content and outbound from the same place. Prospect emails leave from a review-first queue and every send is logged, and Addy, the orchestrator inside TruLata, waits for your OK on anything public.
If you want to know whether ChatGPT, Gemini and the other engines are recommending you today, and what to fix if they are not, open the live demo and see the real dashboard.

™



