Lead generationStrategy, pricing and metrics
Fewer Clicks From Google? Convert Who Still Comes
Google now answers more questions on its own page, so fewer people visit. The ones who still click know what they want. Here is what the research says, and a plan to turn more of those visits into leads.
The short answer
Traffic from Google is falling for many sites because AI Overviews answer questions on the results page. In Pew Research Center’s 2025 study, people clicked a result on 8% of visits with an AI summary, against 15% without one. You cannot win those clicks back, but you can convert more of the visitors who still arrive: answer their real questions on the page, show prices and policies, and let a chat answer and capture contact details at any hour.
The pattern shows up in Search Console the same way for thousands of small businesses. Impressions are flat or rising. Clicks are down. The average position barely moved. Nothing is broken on the site; the searcher simply got an answer before reaching it.
That is the shift Google AI Overviews and AI Mode have brought, and no title tag tweak will undo it. What you can change is what happens after someone clicks. This guide covers what the research measured, who still clicks, and a plan to turn fewer visits into the same number of leads.
What the data says about AI Overviews and clicks
The clearest evidence comes from Pew Research Center, published 22 July 2025. Pew tracked the real Google searches of 900 US adults who shared their browsing data during March 2025. About one search in five (18%) produced an AI summary. When one appeared, people clicked a traditional search result on 8% of visits; without one, they clicked on 15%. They clicked a link inside the summary itself on just 1% of visits. And they were more likely to stop browsing altogether: 26% of visits to a page with an AI summary ended the session, against 16% without one.
Search industry studies point the same way, with different methods. Ahrefs compared 300,000 keywords in desktop Search Console data and found the presence of an AI Overview correlates with a 58% lower average click rate for the top ranked page between December 2023 and December 2025 (published 4 February 2026). Seer Interactive, tracking its own clients’ queries, reported in November 2025 that the organic click rate on informational queries with an AI Overview had fallen to 0.61% by September 2025.
There is a more hopeful 2026 footnote. Seer’s April 2026 update, across 5.47 million queries from 53 brands, found the decline had slowed: the organic click rate on queries with an AI Overview recovered from 1.31% in December 2025 to 2.36% in February 2026. Seer is clear that this is two months of data and that it does not expect click rates to return to pre-AI levels.
Google’s own position is that the picture is better than it looks. In an August 2025 post, Liz Reid, head of Google Search, wrote that total organic click volume was “relatively stable year-over-year” and that Google was sending “slightly more quality clicks”, which it defines as clicks where people do not quickly come back. The post did not publish figures. Google’s Search Central guide to AI features (last updated 10 December 2025) says clicks from results pages with AI Overviews are “higher quality (meaning, users are more likely to spend more time on the site)”.
| Source | What was measured | Finding |
|---|---|---|
| Pew Research Center (July 2025) | 900 US adults, their own Google searches in March 2025 | A traditional result was clicked on 8% of visits with an AI summary, 15% without one. Links inside the summary: 1%. |
| Ahrefs (February 2026) | 300,000 keywords, desktop Search Console data, December 2023 vs December 2025 | An AI Overview correlates with a 58% lower average click rate for the top ranked page. |
| Seer Interactive (November 2025) | 3,119 informational queries, 42 client organisations | Organic click rate with an AI Overview fell to 0.61% in September 2025, from 1.76% in June 2024. |
| Seer Interactive (April 2026) | 5.47 million queries, 53 brands | With an AI Overview, organic click rate recovered from 1.31% (December 2025) to 2.36% (February 2026). Without one: 3.82%. |
| Google (August and December 2025) | Its own statements, no figures published | Total organic clicks "relatively stable year-over-year"; clicks from pages with AI Overviews are "higher quality". |
The honest reading: for questions an AI can answer in a paragraph, fewer people visit, and those who do are, by Google’s account, more engaged. Both point to the same job for your website.
Who still clicks, and why
The AI Overview already gave the searcher a definition and a rough price range. If they still click, the summary did not finish the job, and that tells you something about them.
- They are further along. They know what a boiler service or a dental implant is. They want to know what yours costs, whether you cover their area, and when you can come.
- They have a specific question. “Do you work on Saturdays?” “Can I cancel within 48 hours?” “Do you fix the older model?” General answers do not help them; only your answer does.
- They are comparing. The AI Overview may have named you alongside two competitors. The visitor is opening all three tabs and will act with whichever answers first.
- They arrive at any hour. Search does not keep office hours. A good share of these visits land in the evening or at the weekend, when nobody is at the desk.
A visitor like that is worth more, and is less patient. If the page cannot answer, they try the next tab.
The answer gap: questions your pages do not answer
Most service business websites were written to rank, not to answer. They explain what the service is (which the AI Overview now does) and stop short of the details a buyer needs (which only you can give). Call the difference the answer gap.
You can find yours in an afternoon. Pull the last 50 emails, calls and chat messages from new customers and write down the actual question in each. Then open your website and check whether each one is answered on the page a visitor would land on. In most businesses the same few gaps appear: price, or at least a starting price; the service area; how soon you can come; what happens if they cancel; what is included; and whether you handle the specific case they have.
Every gap is a point where a motivated visitor has to email you and wait, or leave. With 30% fewer visits, each one costs more. This is also the work that helps you appear inside AI answers, as our guide to generative engine optimization explains: specific, current facts on the page are what AI systems quote.
A practical conversion plan
1. Answer the top 20 questions on the page
Take the questions from your answer gap list, rank them by how often they come up, and answer the top 20 on the pages where people land. Not on a separate FAQ page three clicks away: on the service page itself, in plain sentences. A short “Questions we get asked” block at the bottom of each service page works well, and it is easy to keep current.
2. Make prices and policies visible
A missing price sends buyers back to the results page. If your work varies, publish a starting price, a typical range, or what the callout fee covers. Do the same for cancellation, guarantees and payment terms. A visitor who can see the terms can decide today.
3. Add a chat that answers instantly from your own pages
Some questions will always be too specific for a page: “Do you cover the village just past the ring road?” A website AI chat trained on your own pages and documents can answer those in seconds, in the visitor’s words, while they are still on the tab. The important rule is that it answers only from your content and says so when it does not know, then offers a person. Our article on how accurate AI chatbots are covers how to test that before you switch it on.
4. Capture the after-hours visitor
A visitor at 21:30 will not wait until morning for an answer; they will open the next result. A chat that answers at night and asks for a name and email before the conversation turns that visit into a lead your team can call first thing. Our guide to capturing website leads after hours walks through the setup.
5. Let people book inside the chat
For appointment businesses, the shortest path from question to customer is a booked time. If the chat can read your calendar and offer real free slots, the visitor who asked “can you come Thursday?” leaves with a booking instead of a promise to call back.
How to measure it
In Google Search Console, open the Performance report and compare the last three months with the same months a year earlier. Google’s documentation says traffic from AI features is included in this report under the “Web” search type, so it is not shown separately. Look at clicks, impressions and click rate by page. Note which pages lost the most clicks while keeping their impressions: those are usually informational pages where an AI Overview now answers the question.
On your own side, track the number that matters: leads per visit from search. Count form submissions, calls and chats with contact details, and divide by search visits for the same month. If you run a website chat, its reports add four useful numbers:
- Chats started, as a share of visits. This tells you how many visitors had a question the page did not answer.
- Chats handled by the AI, against those passed to a person. A rising share means your content is doing more of the work.
- Leads with an email, from chats where the visitor filled in the pre-chat form.
- The topics it struggles with. This is your answer gap, refreshed as visitors ask.
Review monthly. If leads per visit rises faster than visits fall, the business is fine, whatever the traffic chart says.
Worked example: 30% fewer clicks, the same number of leads
Brookside Plumbing is a firm we made up for this example. A year ago its website drew 2,000 visits a month from Google search, and 2% of them, 40 people, filled in the contact form. Then AI Overviews started answering “why is my boiler losing pressure” and similar questions on the results page. Search visits fell 30%, to 1,400 a month. At the same 2% rate, that is 28 leads: 12 fewer jobs to quote every month.
The owner does three things. She adds starting prices, the service area and the cancellation policy to each service page, which lifts the form rate slightly, to 2.2%. She adds a website AI chat that answers from those pages, with a pre-chat form asking for a name and email. And she connects the chat to her calendar so visitors can book a visit slot.
| A year ago | Now, no changes | Now, with the plan | |
|---|---|---|---|
| Visits from Google search | 2,000 | 1,400 | 1,400 |
| Contact form leads | 40 (2.0%) | 28 (2.0%) | 31 (2.2%) |
| Chats started | 0 | 0 | 70 (5% of visits) |
| Chats with an email, not also a form | 0 | 0 | 9 |
| Leads in the month | 40 | 28 | 40 |
| Visit to lead rate | 2.0% | 2.0% | 2.9% |
The maths: 1,400 visits at 2.2% gives 31 form leads. Five per cent of visitors start a chat, which is 70 chats. Say 9 of those leave an email and did not also fill in the form, many of them in the evening. That brings the month back to 40 leads. The visit to lead rate has gone from 2.0% to 2.9%, and the firm is quoting as many jobs as it did when traffic was higher.
The general rule is simple to work out for your own site. To keep leads flat when visits fall by a share d, your conversion rate has to rise by a factor of 1 ÷ (1 − d). A 30% fall needs a rate about 1.43 times higher: 2.0% becomes 2.86%. A 20% fall needs 1.25 times. That is a far more achievable target than winning the clicks back.
Checklist: make every click count
Measure the drop. Search Console, last three months against the same months a year ago, by page.
Measure your rate. Leads from search divided by visits from search, for the same periods.
Find the answer gap. List the real questions from your last 50 customer enquiries.
Answer the top 20 on the page. On the service pages people land on, in plain sentences.
Show prices and policies. Starting prices, service area, cancellation and what is included.
Add a chat that answers from your pages. With a clear route to a person when it does not know.
Ask for an email before the chat starts. Name and email in a short pre-chat form, so each chat can become a lead.
Cover the evenings and weekends. Let the chat answer and capture contact details out of hours.
Offer booking in the chat. Real free times from your calendar, if you sell appointments.
Review monthly. Visits, leads per visit, chats started, chats handled by AI and unanswered questions.
Where EzyConn fits
EzyConn’s website AI agent is built for the visitor who clicked with a specific question. You give it your website address and it reads your pages; you can add files and write exact answers to the questions that matter most. It answers visitors from that content in multiple languages, and when it does not have the answer it says so and offers to connect them with your team, passing the whole conversation to your inbox or Slack.
- Pre-chat form. You choose the fields: name, email, phone, department, consent. A chat with an email counts as a lead in your reports.
- Booking in the chat. From the Starter plan, the agent offers real appointment times from a connected Google Calendar or Microsoft 365 calendar and books them, with confirmation by email. How bookings work.
- A person when it matters. The visitor accepts the offer, the AI steps back, and a teammate picks up with the full history.
- Reports. Conversations handled by the AI, escalated or led by a person, first reply times, the reasons questions went unanswered, and CSV export for your own spreadsheet.
The Free plan includes the AI agent, needs no card and does not expire, and the website chat goes live in about 3 minutes once you paste one line of code. You can add a free AI chatbot to your website, see how it fits into lead generation, or compare plans on the pricing page. Or simply start free and ask it the five questions your customers ask most.
AI Overviews and falling traffic: common questions
Why are my impressions up but my clicks down?
Your pages are still being shown, often as a source for an AI Overview or further down a longer results page, but more searchers get what they need without visiting. Pew found people clicked a traditional result on 8% of visits when an AI summary appeared, against 15% without one. Rising impressions with falling clicks is the typical pattern for informational queries.
Do AI Overviews reduce traffic for every website?
The effect is strongest on informational questions, the "what is" and "how to" searches an AI summary can answer in a paragraph. Searches with clear buying or booking intent, local searches and brand searches are hit less. Seer Interactive's April 2026 data also showed click rates on AI Overview queries recovering in early 2026, though not to their 2024 level.
What is a zero-click search?
A search that ends without the person clicking through to any website. The answer was on the results page itself: an AI Overview, a featured snippet, a map pack with opening hours, or a knowledge panel. Pew found that a Google visit ended the browsing session 26% of the time when an AI summary appeared, against 16% without one.
How can I tell if AI Overviews are hurting my site?
In Google Search Console, open the Performance report, compare the last three months with the same three months a year earlier, and sort pages by lost clicks. If impressions held while clicks and click rate fell, mostly on question style queries, AI answers are the likely cause. Then check whether leads fell by the same share. Often they fall less.
Can I still get traffic from AI Overviews?
Yes. Pages cited inside an AI Overview tend to earn more clicks than pages that are not. Seer Interactive found cited brands had a 35% higher organic click rate in Q3 2025, while noting stronger brands may simply be cited more. Clear, specific answers, current facts and visible prices make a page easier for an AI to quote.
What is a good visit to lead rate for a service business?
Published averages vary so widely by industry and traffic source that your own baseline is the better guide. Take last quarter's leads divided by visits from search, then track the same number monthly. The goal in a year of falling clicks is simple: lift that rate faster than visits fall.