TL;DR: Google has renamed Search to Ask. AI Mode picks one business to recommend, not ten to rank. It bases that recommendation on how well your information satisfies the customer’s decision process, including concerns they haven’t voiced. Different search pathways recommend different businesses for the same query. The shift is gradual but the direction is clear. Understand your customers, give them what they need, and the AI will recommend you. Ignore them, and it will recommend someone who didn’t.

Customers no longer search Google. They ask Google.

Google has redesigned the search box for the first time in 25 years. The word “Search” has been replaced with “Ask.” Just one little word changed, but such a massive shift in the dynamic.

Ask a question, get a recommendation. Google is steering users towards an experience where it researches, recommends, and eventually acts on the user’s behalf: checking availability, making calls, booking services. In some cases, Google doesn’t even need the user to be specific. It infers what they need and anticipates concerns they haven’t raised.

It goes deeper than that. Personal Intelligence, announced at I/O and rolling out globally, connects Gmail, Google Photos, and soon Calendar. Google will know what you ordered last week, where you went on holiday, what appointments you have coming up. Its recommendations are based on your personal history and context.

Think about what this means for your business. The relationship between your customers and you is increasingly mediated. What to show them, how to frame it, who gets recommended is decided by Google. Your website may no longer be the first touchpoint. But it’s still the touchstone for whether you’re included at all. The process happens inside Google’s interface, on Google’s terms, using information Google has gathered about your business from every source it can find.

This isn’t a feature update. The relationship with your customers is no longer yours to control.

Same question, different answer

There are now two AI layers in Google’s results, and they behave differently.

AI Overviews appear above standard results. They extract facts from web pages and present a summary. Useful, sometimes accurate, occasionally citing the wrong source for the wrong reason. They’ve been spreading steadily from informational queries into commercial and transactional territory. Pre-I/O data shows them triggering on nearly 40% of informational queries, and the rate is climbing across other query types.

AI Mode is a different beast. When a user enters AI Mode, Google doesn’t present a list of options. It picks one business and explains why. It synthesises reviews, evaluates business data, and makes an editorial judgment about who best answers the user’s need. Not who ranks highest. Who best fits.

The finding that should keep you up at night: the same query can recommend different businesses depending on which pathway the user takes. Standard search might show three optometrists equally. AI Mode picks one and doesn’t mention the others. The business that wins in standard search can be absent from AI Mode entirely. You don’t control which path the user takes, and you can’t optimise for one without understanding both.

Google is recommending based on customer needs, not your marketing fluff

AI Mode doesn’t rank pages. There are no positions. There’s a recommendation, and there’s everything else.

Google is drawing on everything it knows and can find about your business, and evaluating whether that information satisfies the customer’s decision process. What does it cost? Can I get it now? How quickly will they respond? Is it safe? What did other people experience?

An optometrist search surfaced bulk billing information the user never asked about. An emergency electrician search led with a safety warning before making any recommendation. Google is anticipating the concerns behind the query, not just matching the words in it.

If your content and reviews address those concerns in specific, concrete terms, you’re in the running. “Premium service” and “trusted provider” give the AI nothing to work with. “Available 24 hours, typical response time under an hour, fully licensed, servicing the eastern suburbs” does. When that’s backed up by reviews mentioning fast callouts and professional service, the AI has the consensus it needs to recommend you.

The businesses getting recommended are the ones whose information most completely satisfies the question the customer is actually trying to resolve.

Content really matters now. Unique, useful, and user-focused, or don’t bother.

If the AI is synthesising answers and making recommendations, why bother creating content at all?

Because the AI has to get its information from somewhere. If you have genuine expertise in your field, the AI needs your input to give good answers. Without fresh, expert content flowing into the system, the AI’s recommendations stagnate and degrade. You’re not just a business hoping to get recommended. You’re part of the supply chain that makes the recommendations worth anything.

But the nature of what you’re creating has to change. Content designed to capture search traffic by targeting keywords and restating what’s already out there is commodity content. Google said as much in their AI optimisation guide. The AI can generate that itself. There’s no reason for it to retrieve a page that does nothing it can’t do on its own.

What works is content that demonstrates what you actually know. Specific expertise applied to real problems. Genuine experience described in enough detail that the AI can extract useful, detailed information from it. If you’re a plumber writing about leaking taps, explain what the homeowner can legally fix themselves in NSW (replacing a washer, yes; touching the water supply, no), when they need to call a professional, what a callout typically costs, and how long a proper repair takes. That’s content the AI can use and the customer can trust.

But write it for the customer, not for yourself. What does the person reading this actually need to know to make a decision? How much is this going to cost me? Can I fix it myself or do I need a professional? How long will it take? What should I expect? When your content answers those user needs in plain language, you’re giving the AI exactly what it needs to recommend you to someone in that situation.

Your content also builds the entity knowledge that shapes how the AI describes and recommends your business. Every piece you publish under your name, linked to your profiles, consistent with your expertise, strengthens the AI’s understanding of who you are and what you know.

Publish with intention. Make it count. And build direct relationships with your customers that don’t depend on Google sending you traffic, because that traffic is no longer guaranteed.

If you need help developing a content strategy that works in this new landscape, get in touch.

The AI future is coming. This is what you need to do now to cope with it.

The shift is gradual. AI Mode is available but opt-in in Australia right now. The design nudges are gentle. Most users are still asking questions and scanning results the way they have for twenty years. Nobody needs to panic today.

But the direction is unmistakable. And the businesses that prepare now will be the ones the AI recommends when the nudges get less gentle.

Start with your content. Every product page, every service description, every piece of information about your business should answer the customer’s decision process in specific terms. What do you do, what does it cost, how quickly can you deliver, what makes you different from the alternative. Not in marketing language. In the language your customers actually use when they’re weighing up their options.

Look at your reviews. Not just the star rating. Are your customers describing their experience in enough detail for the AI to understand what working with you is actually like? Speed of response, quality of outcome, how they were treated. If your reviews are generic (“great service, highly recommend”), they give the AI nothing distinctive to work with. Encourage customers to be specific. The detail in their words shapes the AI’s recommendation.

Check your business information across every surface the AI can access. Your website, your Google Business Profile, your directory listings, your social profiles. If the information is inconsistent, incomplete, or out of date, the AI has to guess. And when it guesses, it may guess wrong, or it may just recommend someone whose data is cleaner.

Think about the third-party picture. What do industry publications, review sites, forums, and social media say about you? The AI draws from all of it. If the only source of information about your business is your own website, the AI has a thin consensus to work from. Earned media, expert mentions, and genuine community participation all contribute to the picture the AI builds of who you are.

And above all, think about your customers. Not about Google. Not about rankings. Not about keywords. About the person asking that question. What do they need to know? What are they worried about? What would make them confident enough to choose you? Answer those questions better than anyone else in your space, and the AI will recommend you.

We help businesses with SEO, generative engine optimisation, and content strategy that positions them for AI search. If any of this resonates, let’s talk.

What Google somehow neglected to mention

Everything above assumes Google’s intentions are good. Maybe they are. But it’s worth noting what wasn’t in the I/O keynote.

Google is positioning itself between your customers and your business. Not alongside. Between. The agentic trajectory means Google mediates the interaction, controls how your business is presented, selects who gets recommended, and eventually handles the transaction. Your customer asks a question. Google provides the answer, the recommendation, and the booking. Your role narrows to fulfilment. And possibly paying a commission for the privilege.

Nobody at I/O mentioned how advertising works in this model. Google’s revenue depends on people clicking ads. Every feature they announced reduces clicks. Information agents deliver answers without search. Agentic booking bypasses websites. AI Mode recommends one business, not a page of options with ads above them. The economics of this haven’t been explained. When a company that makes its money from advertising builds a product that eliminates the need to click, and says nothing about how the money works, the answer is probably worse than the question.

Google has wrapped itself around the content ecosystem like a python. Every feature tightens the squeeze. At some point, the system suffocates. The AI needs fresh, expert, original content to give good recommendations. If ad-supported publishers get less traffic because the AI synthesised their reporting without sending readers, why keep publishing? If review and comparison sites can’t sustain themselves because the AI extracts their findings without sending visitors, the quality of the AI’s recommendations degrades, and the answers get poorer.

Understanding your customer’s needs, creating specific and useful content, building a strong review profile from satisfied customers: all of that is worth doing regardless of Google’s intentions. But do it with your eyes open. Google’s interests and your interests are aligned right now. They may not always be.

Illustration of a python constricting around a glowing sphere of web content, representing how Google's expanding AI features are tightening around the content ecosystem they depend on.

So, would you recommend me?

The businesses that thrive through this shift won’t be the ones that cracked the algorithm. They’ll be the ones that understood their customers well enough that any system, AI or otherwise, would recommend them.

If you found this useful, would you recommend me? Because that’s exactly how it works now.

Jeremy Beckett has been working in search since 2003. Belmore Digital is a SEO and GEO consultancy based in Sydney’s Eastern Suburbs.

This is the third in a series on schema, structured data, and AI search. Read the first post on whether schema markup helps LLMs and the second on Google’s AI guide, the Ahrefs schema study, and the AI content debate.

Frequently asked questions

What is Google AI Mode?

A conversational search experience where Google researches a query, synthesises information from multiple sources, and recommends a single business or answer rather than presenting a list of results.

Is AI Mode the default in Australia?

Not as of May 2026. It’s available in Chrome and Safari mobile on google.com, but users have to actively choose it. Google is nudging users towards it through design choices, particularly on commercial queries. Safari desktop doesn’t have it at all as of May 2026.

How is AI Mode different from AI Overviews?

AI Overviews extract facts and present a summary above standard results. AI Mode goes further: it picks a winner, explains why, and anticipates user needs that weren’t stated in the query.

Can the same query recommend different businesses in AI Mode vs standard search?

Yes. Testing in May 2026 showed different businesses recommended for the same query depending on which pathway the user took. A business can be the top result in standard search and absent from AI Mode, or vice versa.

What makes a business get recommended in AI Mode?

Specific, detailed content that addresses the customer’s decision process. Strong reviews with attribute-specific detail. Accurate and complete business information across every surface the AI can access. The AI is looking for consensus across your own content, reviews, and third-party sources.

Does schema markup help with AI Mode?

Google has said schema isn’t required for AI search but recommends keeping it as part of your overall strategy. It’s infrastructure that supports entity recognition and data accuracy, not a competitive advantage on its own. The earlier posts in this series cover the evidence in detail: Does Schema Markup Help LLMs?

What happens to PPC in AI Mode?

Google hasn’t explained how advertising works in AI Mode. As of May 2026, for urgent, high-intent queries (emergency services, immediate needs), PPC ads still appear and still work. For broader queries, sponsored products are appearing alongside AI answers. The long-term ad model is unclear.

Should I still create content if the AI just synthesises it?

Yes. The AI needs fresh, expert content to give good recommendations. Your content builds the entity knowledge that shapes how the AI describes and recommends you. But create it for your customers, not for search traffic. Unique, useful, and user-focused, or don’t bother.

Illustration of businesses entering a selection funnel with only one emerging as Google AI Mode's recommendation, representing how AI search picks winners rather than listing options.

Get Your Free Copy of JB’s Digital Trends!

This year’s Digital Trends are an absolute cracker. I am seriously pleased with them.

Chock-full of detailed strategic insight, they are an absolutely essential read for any Senior Digital Marketer, C-suite, or Founder who wants to put themselves at the forefront of the game.

Every word, every sentence, is intentional and crafted to within an inch of its meaning.

Know this: I value your time. It won’t be wasted. 🫶

📥 Download your FREE copy today!

JB's Digital Trends 2026