If you run a small business, you have probably noticed that Google looks different lately. There is a conversational tab sitting alongside the usual results, the answers are getting longer, and fewer searches seem to end in a plain list of blue links. That change has a name, AI Mode, and it has grown faster than almost anything Google has launched. Here is what it actually is, what it genuinely changes for your website, and what is worth doing in response.
What is Google AI Mode, and how is it different from AI Overviews?
The two get muddled constantly, so it is worth separating them. AI Overviews are the AI-written summaries that appear at the top of a normal results page for some queries. You search as you always have, and above the traditional results you get a short synthesised answer drawn from across the web. The blue links are still there underneath.
AI Mode is a different thing entirely. It is a full conversational search experience rather than a box bolted onto a results page. You ask a question, you get a synthesised answer, and then you can keep asking follow-ups in the same thread without starting a new search. It behaves much more like a conversation than a lookup.
Both are powered by Google's AI models, but they are separate surfaces with different behaviour. Confusing them leads people to the wrong conclusions (usually either that nothing much has changed, or that search is finished). Neither is true.
How big has AI Mode actually become?
This is the part most business owners have not registered. AI Mode has passed one billion monthly users, and query volume has more than doubled every quarter since it launched. Alongside that, Google has rolled out the biggest redesign of its search box in over twenty-five years. It is now AI-powered and built to accept longer, more descriptive input rather than a couple of keywords.
Redesigning the search box is not a small signal. That box has been more or less untouched for a generation because it is the single most valuable piece of interface on the internet. Changing it means Google expects the way people type questions to change permanently. This has moved well past the pilot stage, and a large share of searching already happens this way.
What actually changes about the way people search
The most practical shift is the length and shape of the query. People used to type two or three words and then sift the results themselves. Now they describe their situation in full sentences, including the constraints and context that matter to them, such as the budget, the deadline, the town they live in, and the thing they have already tried that did not work.
Because follow-ups happen inside the same session, the AI accumulates context before any website gets recommended. Someone might start with a general question about a type of service, narrow it to their area, then ask what a reasonable price looks like, then ask how long the work usually takes. By the time a business is surfaced, the system has a fairly detailed picture of what that person wants. Your content is being assessed against a specific situation rather than a single keyword.
The honest downside: fewer clicks on informational searches
There is no point pretending this is all upside. For purely informational questions, fewer people now click through to any website at all, because the answer is already there in the interface. If someone wants to know what a term means, or how a process works in general, they get that answer and the search ends.
If your traffic is heavily weighted toward broad informational content (general explainers, definitions, lightweight how-to posts), expect it to decline over time regardless of what you do. That is worth planning for honestly rather than discovering in a dashboard six months from now. The more useful question is how much of that traffic was ever going to turn into enquiries in the first place.
Where the opportunity still is
Queries with commercial or local intent still need a real business at the end of them. Nobody gets their bathroom renovated, their accounts filed, or their website built by an AI summary. At some point a person has to choose a provider and contact them, and that step cannot be synthesised away.
So content that helps someone decide and then act retains its value: the pages explaining what you do, who you do it for, what it costs, how long it takes and what happens after they get in touch. If anything, that material matters more now, because it is what an AI system draws on when it is asked to recommend someone.
Tip
What to actually do about it
None of what follows is exotic. It is mostly ordinary good practice applied with more discipline than most sites manage.
Answer specific questions directly and early
AI systems extract passages that clearly answer a question. If the answer to "how long does this take" is buried in the fourth paragraph behind a paragraph of scene-setting, it is far less likely to be pulled out and used. Put the direct answer near the top of the relevant section, in plain language, and then expand on it. This also happens to be better for the humans reading the page.
Include concrete specifics
Name the areas you serve. Describe the steps of your process. Give realistic timeframes and some pricing context, even if it is a range or a starting figure with the variables explained. Vague marketing copy gives an AI nothing to cite, and a line like "bespoke solutions tailored to your needs" is not information. Specifics are what get quoted.
Keep pages current
Outdated information is less likely to be surfaced, and more likely to do damage if it is. Prices that have changed, services you no longer offer and a team page listing someone who left two years ago all undermine the impression of a business that is actually operating. Reviewing your main pages once or twice a year is usually enough.
Make sure your content is server-rendered
This is the one genuinely technical point, and it catches out a lot of otherwise well-built sites. AI crawlers generally do not execute client-side JavaScript. A page that only assembles itself once it loads in the browser can be effectively invisible to them. A human sees a full page, while the crawler sees an empty shell. If your site was built as a single-page application without server rendering, this is worth checking before anything else on this list.
Info
Strengthen the fundamentals
Clear structure, descriptive headings, and genuine expertise written out properly. These help AI extraction and human readers in exactly the same way, which is a reasonable sign that they are worth doing regardless of what any particular platform does next. Most of my SEO work on small business sites starts with exactly these checks.
What not to bother with
Be wary of anything marketed as an AI-specific trick, or tooling that promises guaranteed citations in AI answers. There is no verified mechanism for buying your way into an AI response, and the people selling one are usually selling confidence rather than results.
The practices that work here are the ones that have always worked, which means clear, specific, current, technically sound content from a business that knows its subject. What has changed is how little tolerance there now is for doing them badly.
Where this leaves you
AI Mode is a real shift, and some traffic patterns will not return to what they were. But the part of search that matters most to a small business (somebody with a problem, a budget and an intention to hire) still ends with a person choosing a provider. Make it easy for them, and for the systems now doing the introductions, to understand exactly what you do and who you do it for. That is the whole job, and it has not changed as much as the headlines suggest.
I work on small business sites so the pages that win enquiries are specific, current and readable by both people and AI crawlers. Send the project brief with your goals and the pages you want looked at.
