How do we win back the traffic we are losing to Google AI Overviews?

“Who can help us win back traffic we’re losing to Google AI Overviews?”
Many marketing teams see the same pattern. Rankings hold. Impressions hold. Clicks fall. Online stores often feel it first, because a large part of their traffic used to come from searches that Google now answers at the top of the page.
This article is about what you can still win, and how. It starts with the part many guides skip: some of those clicks are not coming back.
The short answer: five steps
Avisible recommends this order:
- Accept that some clicks will not come back. Aim to be named, to be described correctly, and to get the click when the buyer is ready to buy.
- Keep doing SEO. According to Ahrefs, 76% of the URLs cited in Google AI Overviews also rank in Google’s top 10. For ChatGPT the figure is 8%, and for Gemini 8.6%.
- Put product facts in text, and keep prices and shipping terms the same everywhere. AI cannot quote what it cannot read, and it repeats what resellers say about you.
- Earn reviews and cover everyday use, not only your hero product. AI gives brands lanes, and it keeps you in yours.
- Measure what happens inside the answer. Track how often open questions name you, per AI system, and whether the facts are right.
Some clicks will not come back
The research points in one direction. Pew Research Center (July 2025) found that Google users clicked a result in 8% of visits where an AI summary appeared, against 15% of visits without one. Ahrefs (February 2026, 300,000 keywords) found that when an AI Overview appears, the click-through rate of the first result is on average 58% lower.
For simple questions, the summary is the answer. A question like “what size do I need?” no longer needs a visit to your store. If your goal is to win all of that traffic back, the goal is wrong.
A better question is this: when a buyer asks AI for a recommendation, is your store named, is it described correctly, and do you get the click when the buyer is ready to buy?
The new goal: be named, be right, get the ready buyer
- Named. Your brand or product appears in answers to open questions, where the buyer has not typed your name. In Avisible’s measurements, almost every brand is named in 90–100% of answers when the question includes its name, but only in 0–40% of answers to open questions. Open questions are where new customers come from.
- Right. Prices, shipping terms, product names and use cases in the answer match your own facts.
- Clicked when ready. At the purchase step, the answer links to you, or to a store where your price and terms are correct.
Google AI Overviews still lean on your SEO
AI Overviews and chat assistants do not read the web in the same way. Ahrefs (August 2025, 15,000 queries) found that 76% of the URLs cited in AI Overviews also rank in Google’s top 10. For ChatGPT the figure was 8%, and for Gemini 8.6%.
This has two consequences. First, SEO still matters for AI Overviews. The pages that rank well are the pages Google’s summary draws on, so keep doing that work. Second, good rankings alone do not get you into ChatGPT or Gemini. They often read other sources: marketplaces, review sites, forums, reseller pages and your own product text.
The engines also behave differently over time. In one Avisible audit, Google AI Overviews gave the same text in all three rounds for 94 of 95 questions, while ChatGPT and Gemini changed from round to round. One check tells you a lot about AI Overviews. One check of ChatGPT tells you much less.
What online stores should fix first
In e-commerce, most of the lost visibility in AI answers comes from a few practical gaps. These are the ones Avisible sees most often.
- Put product facts in text. Materials, sizes, use cases, warranty and care instructions belong on the product page as plain text. One consumer brand had key product information in images without alt text, so AI could not read it.
- Keep prices and shipping the same everywhere. A consumer brand had two online stores that disagreed on shipping terms. AI quoted a price three times higher in the US and gave the wrong shipping terms. Compare your own stores, country sites and product feeds with each other.
- Fix reseller and marketplace listings. For one industrial product, 67 of 71 buying answers sent the buyer to a reseller, and the maker’s own store price appeared in 0 of 448 answers. In 15 Gemini answers the product appeared without a name, because a reseller’s listing title left out the model name. Your resellers’ pages are part of your AI visibility.
- Watch the shopping panels. For the same manufacturer, Gemini relied on Google shopping panels in 98 answers. Keep your product feed and shopping listings complete and correct, because that is the data these panels show.
- Earn reviews and discussion. For one consumer brand, Reddit was the most-cited source in English answers. Real customer discussion shapes what AI says about you. You cannot write it yourself, but you can make reviewing easy and answer questions where customers ask them.
- Cover everyday use, not only your hero use. AI gives brands lanes. One consumer brand won extreme-weather questions in 83–100% of answers, but was named in only 0–17% of everyday-use questions. If your pages talk about one use only, AI keeps you in that lane.
Measure the new KPIs
Traffic and rankings still matter, but they do not show what happens inside the answer. Add these to your reporting:
- The share of answers to open questions that name your brand or product, per question.
- The same figure per engine. Google AI Overviews, ChatGPT and Gemini can differ a lot.
- The competitor named most often in the questions you lose.
- Whether prices, shipping terms and product facts in the answers are correct.
- The sources the answers cite: your site, resellers, marketplaces, reviews, forums.
Measure with care. AI answers vary, so ask each question several times. Decide how you count empty answers: in one audit, the same data read 87% or 88.5% depending on whether empty Google answers were counted. You can track a few questions by hand in a spreadsheet. For more, you need a tool that repeats the questions every week.
Who can help
Avisible helps online stores and consumer brands win back the part of AI search that can still be won. Avisible is an AI visibility platform and service: it measures how ChatGPT, Gemini and Google AI Overviews name and describe your brand, and shows what to fix first.
Avisible offers both a one-off Snapshot or Audit, and continuous tracking on the Avisible platform.
- Visibility Snapshot (€499). 100 buyer questions × 3 rounds × 3 AI systems (ChatGPT, Gemini, Google AI Overviews) = 900 AI queries, delivered in three working days. A one-off report, no subscription.
- Visibility Audit. 200+ buyer questions built for your category and approved by you, six AI systems, competitor benchmarking, source mapping, site analysis and a prioritised 90-day action plan. Four weeks end to end, about 4 hours of your time, typically €5,000–25,000.
- Platform. A tool for continuous tracking. It follows your questions weekly and shows the share of answers that name your brand, by topic, and the competitor named most per question.
Avisible works alongside your SEO agency, not instead of it. Your agency keeps improving rankings, which feed AI Overviews. Avisible shows what happens inside the answers, including ChatGPT and Gemini, and which sources and listings need work.
Method in brief
The client figures come from Avisible Snapshots and Audits from October 2025 to September 2026 for consumer brands, e-commerce and industrial companies. Each question was asked in several rounds in each AI system, and clients are not named. The external figures are from Pew Research Center and Ahrefs.
Read also
Avisible measures how ChatGPT, Gemini, Perplexity and Google AI Overviews describe and recommend brands, from retail and hospitality to B2B services, including in the Nordic languages. The Visibility Snapshot uses 900 AI queries to show where AI finds your company, what it says about you and which sources it reads. Delivered in three working days, €499.