What Google's AI Mode Actually Does to Search Results
Google has made AI Mode the default experience for search in the United States, and it is rolling out to more markets quickly. This is not a minor interface tweak. It fundamentally changes what users see when they type a query. Instead of a list of blue links topped by a featured snippet, most queries now return a synthesised AI-generated answer at the top of the page, with source citations tucked in below or to the side.
The practical effect is stark. A user searching "best noise-cancelling headphones under £150" no longer necessarily scrolls through product pages. They read a paragraph written by Google's AI, click one or two of the cited sources if they want more detail, and then make a decision. The traditional first-page ranking is no longer the finish line. Being cited inside the AI answer is.
This shift has been building for a while. AI Overviews (the feature previously known as Search Generative Experience) were the trial run. AI Mode is the full deployment. Google is betting that users prefer a conversational, summarised answer over a page of links, and early engagement data suggests they are right. For businesses, that means the old playbook needs updating urgently.
Why Organic Click-Through Rates Are Falling Even for Ranked Pages
One of the most disorienting things about AI Mode is that your rankings may stay exactly where they are, but your traffic drops anyway. This is already happening. Pages sitting at position one for competitive queries are reporting double-digit percentage falls in organic clicks because the AI answer satisfies the user before they reach the link.
Think about what that means structurally. You invested in content, earned backlinks, achieved a top-three ranking, and now a large proportion of the users who see your result never actually visit your site. Google has essentially monetised your effort by feeding it into its AI and keeping the user on Google.
The businesses that are holding up best under this pressure share a common characteristic: they appear inside the AI-generated answer itself, as a cited source. That citation is the new click. It carries authority signals to the user, it often does generate a visit, and it positions your brand as trustworthy in the eyes of someone who is already in a buying mindset.
Getting cited is not random. AI Mode, like ChatGPT and Perplexity, draws on structured signals when deciding which sources to reference. Pages with clear, machine-readable content consistently outperform pages where the information is buried in dense paragraphs with no structural signals. This is where schema markup becomes directly relevant to your revenue.
The Signals Google's AI Uses to Choose Its Sources
Google has not published a precise ranking formula for AI Mode citations, but the patterns are clear enough from testing and observation. There are three categories of signal that appear to matter most.
Entity clarity
The AI needs to understand who you are and what you do without ambiguity. If your site does not explicitly define your brand, your product categories, your location, and your area of expertise in a structured way, the AI treats you as a fuzzy signal. Fuzzy signals do not get cited. Schema types like Organization, Brand, and WebPage help establish entity clarity. You can read more about the latter in our guide on how to use WebPage schema to help AI search understand your site.
Topical authority and depth
AI Mode cites sources that go deep on a topic, not those that try to cover everything shallowly. A 600-word product description will rarely be cited in an AI answer about that product category. A detailed page that answers the real questions buyers ask, includes specifics like materials, dimensions, compatibility, and return conditions, and uses structured data to make that information machine-readable, stands a much better chance.
Trust and verification signals
Reviews, credentials, return policies, and pricing transparency all feed into how trustworthy a source appears to an AI model. Schema types like AggregateRating, MerchantReturnPolicy, and Offer encode exactly this kind of trust information in a format the AI can parse reliably. These are not cosmetic additions. They are functional signals that directly influence citation decisions. Our post on using AggregateRating schema to influence AI recommendations goes into the mechanics in detail.
Practical Steps to Take Right Now
Knowing that AI Mode is the default is one thing. Doing something about it is another. Here is a prioritised action list based on what actually moves the needle.
1. Audit your current structured data coverage
Most e-commerce sites have basic Product schema and not much else. Run a structured data audit and identify gaps. Are your reviews marked up? Is your return policy encoded? Do your product pages include Offer schema with price, availability, and currency? If you are not sure where to start, FlinnSchema's free AI visibility audit is a practical first step that maps your current schema coverage against what AI search engines are actually looking for.
2. Add structured data to your highest-traffic pages first
Do not try to fix everything at once. Prioritise the pages that already attract the most organic traffic, because those are the pages that have the best chance of appearing in AI-generated answers. Add or improve schema markup on those pages, then work down the list. A product category page with 10,000 monthly visitors is worth far more attention than a blog post getting 50.
3. Rewrite thin content to answer real questions
AI Mode pulls from pages that directly answer what users are actually asking. Look at the questions appearing in "People Also Ask" for your key queries, check forums like Reddit and Quora in your niche, and use those to identify the gaps in your content. A product page that answers "does this work with X?" or "how long does shipping take?" in clear, on-page text is more likely to be cited than one that stops at the product description.
4. Encode your trust signals explicitly
Return policies, guarantees, certifications, and review counts should all be present in your schema markup, not just visible on the page. AI models cannot always reliably extract information from free-form text. Give them structured data and they will use it. Our guide on MerchantReturnPolicy schema for AI shopping trust walks through the exact properties to include.
5. Think about brand disambiguation
If your brand name is shared with another business, a common word, or a person, the AI may be citing the wrong entity or ignoring yours entirely. Brand schema and Organisation schema with sameAs properties pointing to your Wikidata entry, LinkedIn, and social profiles help AI models pin down exactly who you are. This sounds like a detail. It is not. Brand confusion is one of the most common and most damaging issues we see in AI visibility audits.
What This Means Specifically for E-commerce Brands
E-commerce brands are in the most exposed position right now. Google Shopping has always been a paid channel, but organic product discovery was a meaningful traffic source for many stores. AI Mode is compressing that organic opportunity significantly.
The brands that will weather this well are those that appear in AI answers for shopping queries. To do that, your product pages need complete Offer schema (price, availability, currency, shipping details), AggregateRating markup pulling in your review data, and MerchantReturnPolicy schema that tells the AI exactly what your return window and conditions are. Together, these signals make your products legible to AI shopping recommendations in a way that plain HTML pages simply are not.
Shopify stores in particular have a schema gap problem. Shopify's default theme outputs some basic Product schema, but it frequently omits return policies, review aggregates, and shipping details. Closing those gaps manually or through a structured data solution is one of the highest-ROI actions a Shopify merchant can take right now.
The Longer Game: Building AI Share of Voice
One framing shift that helps here is to stop thinking purely about rankings and start thinking about share of voice in AI answers. For any given topic your business owns, how often does an AI model cite your brand, your products, or your content? That is your AI share of voice, and it is measurable.
Tracking this means running regular prompt tests across ChatGPT, Perplexity, and Google AI Mode for your key queries and recording which sources appear. It is time-consuming to do manually, but it reveals patterns quickly. You will find that some competitors are being cited far more than their actual quality warrants, simply because their structured data is cleaner and their content is better organised.
Closing that gap is a structured, systematic process. It involves schema implementation, content improvements, and ongoing monitoring. It is not a one-time fix. But every week you delay is a week your competitors are building their AI citations while yours stagnate.
If you want to understand where your brand stands today, the practical starting point is an audit. FlinnSchema's free AI visibility audit gives you a clear picture of your current structured data, your citation patterns across AI engines, and the specific gaps to address first. From there, the path forward is a lot clearer than it looks from the outside.
Frequently Asked Questions
Will Google AI Mode affect my paid search campaigns?
Yes, indirectly. AI Mode places the AI-generated answer above paid ads in many query types, which pushes ads further down the page. Click-through rates on paid search are already showing pressure in categories where AI Mode delivers a strong answer. This does not mean paid search is dead, but it does mean the economics are shifting. Brands relying entirely on paid traffic need to think about AI visibility as a complementary channel.
Does schema markup directly get you into AI Mode answers?
Schema markup does not guarantee a citation, but it significantly improves your chances. Structured data makes your content machine-readable, which is exactly what AI models need to extract and cite information reliably. Pages with complete, accurate schema markup consistently appear in AI-generated answers more often than equivalent pages without it. Think of it as removing friction between your content and the AI's understanding of it.
How quickly can I expect to see results after adding schema markup?
Google typically re-crawls and re-indexes updated pages within days to a couple of weeks for active sites. However, changes in AI citation patterns can take a little longer to stabilise, sometimes four to eight weeks, because the AI's understanding of your site updates as new crawl data is incorporated. The more authority your domain already has, the faster you tend to see movement.
Is this only a problem for large e-commerce sites, or do small businesses need to worry too?
Small and mid-sized businesses are often more affected, not less. Large brands have enough ambient authority that AI models cite them by default. Smaller brands need to work harder to establish that entity clarity and trust signal footprint. The good news is that structured data implementation is equally available to a small Shopify store as it is to a major retailer. The playing field is more level than it looks, provided you do the work.