AI Search News, July 2026: What Changed and What It Means for UK Businesses

Every month we take stock of where AI search is heading and what it actually means for the kind of independent UK business we work with. This is not a leaderboard of viral news stories, and we are not going to hand you invented statistics dressed up as certainty. It is a straightforward, honest read of the direction things are moving in, written for someone who runs a business and does not have time to follow AI search full time. If you want the deeper, sourced data behind the wider shift, our state of AI search in the UK piece is the place for that. This one is shorter, plainer, and focused on what changed recently and what to do about it.

The big picture: search is answering, not just linking

The theme has not changed this month, it has simply kept building. Across the major platforms, the pattern is the same: a question goes in, and an answer comes back, often with no click required at all. Google increasingly shows an AI-written summary above its traditional results for a growing share of searches. ChatGPT has its own search mode built in, so a user can ask a question and get a synthesised answer with sources rather than a page of links. Gemini sits inside Google's core products and answers directly rather than sending someone off to browse. Perplexity was built from day one as an answer engine rather than a search engine, and it continues to grow as a habit for people who want a summarised answer with citations rather than ten blue links to sift through.

None of this is a single dramatic announcement. It is a direction of travel that has been visible for a couple of years now and is still accelerating. The practical effect for a business owner is simple: the moment where a potential customer decides who to trust is increasingly happening inside an AI answer, not on a results page you can see and optimise the old way.

What's actually shifted lately, in plain English

Here is a general summary of where each of the big four sits at the moment, deliberately kept broad rather than pinned to specific dates or numbers we cannot independently verify.

ChatGPT. Search inside ChatGPT keeps maturing as a genuine alternative to typing a query into a search engine, particularly for people who want a synthesised answer with sources rather than a list to click through themselves. The trend is that more everyday, commercial questions, the kind a customer might ask before choosing a business, are being answered this way.

Google's AI Overviews and AI Mode. Google continues to expand where an AI-generated answer sits above the traditional organic results, and to expand the more conversational AI Mode experience within Search. The direction is consistent: more of the answer is delivered on the results page itself, before a user ever reaches a website.

Gemini. Google keeps folding Gemini deeper into its everyday products, from Search to Workspace to Android, which widens the surface area where an AI answer, rather than a webpage, is the first thing a person sees.

Perplexity. Perplexity continues to grow its user base as a dedicated answer engine, and it remains a useful bellwether because, unlike Google, it is not weighed down by two decades of traditional search behaviour. What it cites today is often a reasonable preview of where the wider market is heading.

We are being deliberately cautious with language here. Anyone telling you an exact percentage of searches now show an AI answer, or naming a specific date something changed, without a citable, checkable source, is guessing or selling. What we can say with confidence is the direction, because it has been consistent for a long stretch: more answers, delivered earlier, with fewer clicks required to get one.

What hasn't changed: the technical basics AI still relies on

Here is the part that gets lost in all the noise about which chatbot is winning this month. Every one of these systems, ChatGPT, Google's AI answers, Gemini and Perplexity, still has to find, read and understand your website before it can recommend you. That has not changed and, on current evidence, is not going to. If anything, it matters more now, not less, because there are more systems reading your site and fewer chances to correct a wrong first impression once an AI has formed one.

The businesses getting named in AI answers tend to share the same unglamorous traits: clean structure, working schema markup, a site that AI crawlers can actually reach, and pages that answer real questions clearly rather than burying the answer in marketing copy. None of that is new advice. It is the same advice that mattered in 2024, and it is the same advice that will matter in 2027. What changes month to month is which platform is growing fastest, not what the underlying technical work looks like.

What UK businesses should actually do about it

If you run a business and want a sensible response to all of this, it comes down to a short, practical list rather than a scramble to chase every new feature announcement.

1. Check your site is actually readable by AI crawlers. A page that reads clearly to a person can still be invisible to an AI system if the underlying markup is a mess, or if a robots.txt file or a security setting is quietly blocking the crawlers these systems use. This is worth checking on a proper cycle, not once and forgotten. We wrote a fuller walkthrough in our piece on AI crawler access.
2. Get your schema and structured data in order. Schema is how you tell Google and AI systems exactly what you do, who you serve and why you are credible, in a format machines can parse reliably rather than infer. Most small business sites have little or none of it.
3. Structure your pages to answer real questions. Headings, clear sections and content that directly answers the questions a customer would actually ask are far easier for an AI system to lift and cite than a page of vague marketing language.
4. Keep your facts consistent everywhere you appear online. AI systems build confidence in a business from consistent, corroborated information across multiple sources. Inconsistent details between your website and other places you are listed make it harder for a system to trust what it finds.
5. Check what AI actually says about you, regularly. Ranking well on Google no longer tells you how you are represented in an AI answer. The only way to know is to ask the tools yourself, on a proper schedule, and adjust based on what comes back.

This is precisely the kind of work we do for clients month in, month out: the technical and schema foundations that make a site genuinely readable to Google and AI, and the citations and authority work that builds the consistent, machine-verifiable presence these systems increasingly rely on to decide who to trust. Our complete guide to AI search optimisation sets out the fuller method if you want to go deeper.

What we're watching next

We are not going to pretend to predict specific announcements. What we are watching, in general terms, is whether AI answers keep expanding into more everyday, local and commercial questions, whether the gap between ranking on Google and being cited by AI keeps behaving the way it has, and whether crawler access keeps tightening as more platforms adjust their defaults. None of that requires a crystal ball. It requires checking the fundamentals regularly and adjusting calmly, which is a far more reliable strategy than reacting to every headline.

The bottom line

Nothing here should cause panic, and nothing here is a reason to buy a guarantee from anyone, because no one honest can promise a specific placement inside a system they do not control. The sensible response is steady, technical and ongoing: keep your site crawlable, keep your schema and structure in order, keep your facts consistent, and keep checking what the AI systems actually say about you. That is unglamorous work, and it is exactly the work that compounds. Businesses that keep this foundation solid month after month tend to be the ones AI systems have learned to trust when the moment to recommend someone arrives.

Want this handled for you rather than added to your own to-do list? Stagg Studios does the technical foundations, structure and citations behind AI visibility every month, from £100/month, with a short plain-English report each time so you always know what changed. See our pricing or get started here.