The UK Just Cracked Google's Search Monopoly: What It Means for You

On Wednesday 23 September, the UK's competition regulator published a document that should matter more to your website traffic than any Google algorithm update this year. You almost certainly didn't see it. That's the point — and that's the problem.

The Competition and Markets Authority (CMA) strengthened its proposals for how Google must present search choices to British Android and Chrome users. The headline: Google would have to show a choice of search providers when someone first sets up an Android phone or opens Chrome, and then prompt them once a year to pick their default again. And — this is the part that's new — AI assistants such as ChatGPT and Perplexity would be allowed on those screens, provided they meet technical and security criteria.

Read that again. A major regulator is treating an AI assistant as a substitute for a search engine, and demanding that users get to pick one as their default. That is not a product announcement. That is a reshaping of the funnel through which most of the planet finds websites.

We've spent the last year advising clients on how to get cited by AI search. This story is the other side of the same coin: the moment AI assistants stop being something people open deliberately, and start being the place they search from by default.

What the CMA actually proposed

Strip away the regulatory language and there are four concrete requirements on the table.

1. A choice screen at setup. When someone first sets up an Android phone or opens Chrome, Google must present a choice of search providers — not just bury alternatives in settings.

2. A yearly prompt. Once a year, users get nudged to pick their default search provider again. Google hates this part. In its March response, the company argued yearly prompts would annoy users and proposed a permanent switch in device settings instead. The CMA kept the yearly prompt anyway. It understands something Google would rather you forgot: defaults win, and re-asking is the only thing that beats a default.

3. AI assistants become eligible. This is the real change. Previously, the eligibility rules effectively barred AI-native tools. The updated proposal deletes the old 'core and central' marketing criterion that kept them out, and opens the screen to any assistant that meets the CMA's technical and security requirements. ChatGPT, Perplexity and whatever launches next year would qualify. The CMA's chief executive, Sarah Cardell, said it plainly: "We've listened carefully to feedback and are responding quickly to the rise of AI assistants, which people are increasingly using as their search service, but aren't included on Google's choice screens."

4. Fair attribution for publishers. Any provider appearing on the screens must fairly attribute publisher content — credit the source and let users reach the original material. This clause exists because publishers screamed during consultation. The Professional Publishers Association argued AI assistants should be excluded entirely, saying they sit outside existing licensing frameworks; the News Media Association warned of traffic being diverted away from publishers. The CMA's answer was a compromise: assistants can be on the screen, but they must show where their answers come from.

The consultation runs until 9 October 2026, with a final binding decision expected by the end of the year. Implementation would follow.

Why the CMA is doing this now

None of this came from nowhere. There is a timeline, and it tells you which way the wind is blowing.

See the pattern. Every six to eight weeks, the regulatory walls around Google's default position get a little higher. The UK is doing what regulators do: start with the designation, add transparency, add publisher rights, then come for the default itself.

And the reason the CMA had to widen the proposal in May is the reason this matters to you. Google's own products were erasing the distinction between "search engine" and "AI assistant". You can't write rules that pretend ChatGPT is a different product from search when Gemini answers are already inside Google Search. Sarah Cardell's other line this week deserves attention too: "This is about future-proofing the options available to people in the UK so they can choose the search service they want as these come to market." Future-proofing. She's saying the category is still being invented, and the rules need to fit things that don't exist yet.

Google built the CMA's argument for it

There is an irony here worth naming. The reason the CMA had to widen its January proposal in May is that Google's own products erased the line the proposal was built on. The May 2026 Gemini 3.5 Flash redesign folded assistant-style answers so deeply into Search that the distinction between "search engine" and "AI assistant" became a fiction maintained only in Google's own policy documents.

Google, in other words, argued itself into this corner. You cannot tell a regulator that ChatGPT is a fundamentally different product from search — and therefore ineligible for the choice screen — while your own flagship product demonstrates daily that search and AI assistance are the same activity. The CMA simply took Google's product roadmap at its word. If Gemini-in-Search is still "search", then ChatGPT is search too, and it gets a slot.

This is also why Google's fight against the yearly prompt matters. The company ships Gemini tightly integrated with Search on the default surface, with no alternative on any choice screen today. Its March argument — that a yearly prompt would "annoy users" and a settings toggle would suffice — reads differently once you see it for what it is: a defence of the last unexamined default in the business. Every other layer of Google's search privilege has now been regulated. The default itself is the final layer, and the CMA is coming for it with a calendar.

The uncomfortable maths of defaults

Here's why a choice screen matters more than most founders realise. Defaults are the single most powerful force in consumer technology. The vast majority of users never change a default — not because they love it, but because it works and they have other things to do. Google has paid tens of billions of dollars for exactly this reason to stay the default on other companies' devices. A default is worth real money precisely because people don't touch it.

So what changes when the default is up for grabs once a year, and the options include ChatGPT?

First, understand the scale of the surface. Chrome holds close to 69% of global browser market share and runs on roughly 3.83 billion devices, according to industry estimates. Android is the dominant mobile OS in the UK and across most of the world, including Pakistan. The choice screen Google shows on that surface is the highest-leverage distribution slot in the search business. Whoever gets a slot there gets a share of the billions of searches that currently flow to Google without a second thought.

Second, look at the mechanics the CMA sketched. The screen works on a "5+7" system: the five most popular providers get slots, plus seven more chosen at random from eligible services. That random-draw half is deliberate — it gives smaller challengers, including newer AI assistants, a genuine shot at screen time independent of their current traffic. This is not a token gesture; it's a funnel.

Third, look at who's already on the screen and who isn't. Gemini ships tightly integrated with Google Search and sits on the default surface. ChatGPT and Perplexity have had no slot on the Android or Chrome default surface at all. The proposal doesn't punish Google's assistant — it just lets the others stand next to it. If you believe your product is better, this is the kind of competition you should welcome. Google's resistance tells you what it thinks.

Our position: the yearly prompt is the single most important line in the whole proposal. A one-time choice screen at setup is easy to game — the incumbent optimises the wording, the ordering, the friction, and the moment passes. A yearly re-prompt turns the default into a subscription that has to be re-earned. That is genuinely threatening to a company whose business model depends on nobody ever reconsidering. Google's complaint that it would "annoy users" is honest, in a way: it would annoy Google most of all.

The publisher clause is the sleeper story

Everyone will talk about ChatGPT on the choice screen. Fewer people will notice the fourth requirement, and it's arguably the bigger deal for website owners.

"Fair attribution of publisher content" sounds like regulatory wallpaper. It isn't. It is the first time a major regulator has said, in the context of defaults, that whoever shows up as a search option must credit sources. AI assistants have spent two years being accused of hoovering up publisher content and answering without attribution. The CMA's fix is blunt: if you want the default slot, you show where the answer came from and you link it.

Combine this with two things that happened in June — publishers gaining opt-outs from AI results, and ranking-transparency rules — and a picture emerges. The UK is building, piece by piece, the first complete regulatory framework for the AI-search era: you can opt out of training/answers, you can see how rankings work, and the assistants that answer must credit you. Whether you think that's fair to publishers or not, it is now the direction of travel, and it is moving fast.

For businesses, the implication is direct. The question we've been asking clients since spring — "can an AI assistant see you, and does it know what you do?" — is becoming a question regulators are asking too. If assistants must attribute sources, then being a citable source becomes the unit of traffic. Vague marketing copy doesn't get attributed. Specific, structured, quotable facts do. This is exactly why we've been pushing structured data and clear entity information as the core of AI-search readiness: when attribution is mandatory, the pages that are easiest to attribute win.

One caution, though — and this is a judgement call, not a prediction. Mandatory attribution could easily turn into a compliance checkbox: a tiny link nobody clicks, like cookie banners. If the attribution is buried, publishers win nothing and the assistants keep the traffic. The CMA's wording is about "fair" attribution, and how strictly that gets enforced will decide whether this clause matters. We'll be watching the final decision text for exactly that word.

What this means for your traffic

Let's be concrete about who should care and why.

If you rely on Google search traffic — most businesses reading this — the honest read is: this is a slow leak, not a dam break. UK users get a yearly nudge; some fraction pick ChatGPT or Perplexity; over a few years a meaningful share of "search" volume migrates into assistants. Your Google rankings won't collapse in Q1 2027. But the trend line on organic sessions bends, and the bend is permanent. The founders who notice early get to build their assistant-channel presence cheaply, while everyone else is still refreshing Search Console.

If you sell in Pakistan or the US — where these rules don't directly apply — you should still treat this as a preview. Regulators copy each other shamelessly, and the EU already has choice screens under the Digital Markets Act. The UK is simply the first to say the quiet part out loud: an AI assistant is a search engine now. When the US Department of Justice's remedies phase against Google was grinding through this same question, the same logic kept appearing. The direction is set; only the timing varies by country.

If you're a publisher or content-led business — the June opt-out rights plus September's attribution clause give you more leverage than you've had in years. The mistake would be using it bluntly. We've written about blocking AI bots and the trade-offs are real: block everything and you disappear from the assistants your future customers are asking. The UK framework rewards a finer strategy — stay visible, stay attributable, and let the rules force the assistants to send you credit.

Here's how the three channels now stack up for a business website:

That last row is the one to sit with. In the assistant-as-default world, a large share of your "traffic" is people seeing your name in an answer without ever visiting. That sounds like a loss, and in pure session terms it is. But it is also a moat: the assistant cites the sources it trusts, and trust compounds. Brands that become the default citation for their category inside ChatGPT and Perplexity will be very hard to dislodge, the same way the #1 organic result is today.

What this looks like from Pakistan

None of these rules apply in Pakistan. The Competition Commission of Pakistan has no equivalent proceeding, and nobody is about to see a choice screen on a phone bought in Lahore. So should Pakistani businesses care?

Yes — arguably more than the British ones. Here's why.

Pakistan's internet runs on Android. Budget and mid-range Android phones are the default computing device for hundreds of millions of people across South Asia, and on those phones Google Search is not a choice anyone ever made; it is simply where the search box goes. The CMA's proposal doesn't change that in Karachi tomorrow. But it changes the product that ships globally. When Google is forced to build choice-screen infrastructure for the UK, the marginal cost of shipping it elsewhere collapses. And when ChatGPT and Perplexity have to meet "technical and security criteria" for a regulator, they build to the strictest standard and ship it everywhere.

There's a second, more immediate angle. Pakistani businesses increasingly sell into regulated markets. If your customers are in the UK or the US — software exporters, freelancers, e-commerce sellers — your customers' search behaviour is what the CMA is reshaping. A Lahore-based agency whose leads come from Google UK searches needs the assistant-channel strategy even more urgently than a London one, because it can't walk into a client's office and explain what happened to the funnel.

And there's a third angle that almost nobody is pricing in: language. AI assistants handle Urdu and Roman Urdu search queries far more naturally than keyword-based Google ever did. A user typing "Lahore mein best web development company kaunsi hai" gets a conversational answer, not ten blue links. As assistants inch toward default status, businesses with strong Urdu-language content and clean entity data will have an advantage that traditional SEO — built for English keywords — never gave them. We have started advising clients to treat Roman Urdu FAQ content as a first-class asset, not a translation afterthought. In our observed range, a bilingual FAQ layer adds roughly 20–30% to a content build, and it's the cheapest assistant-visibility win we know.

What we're telling our clients this quarter

We don't do vague advice, so here's the actual checklist we've been running through with clients since the proposal dropped. It takes most sites two to four weeks, and for a typical small-business site our studio charges in the $800–$2,500 range for the full pass — roughly Rs. 225,000 to Rs. 700,000 at current rates, depending on how messy the existing content is.

1. Get assistant-visible before the assistants get default status. This is the same work we laid out in our AI search playbook: schema markup on every money page, a clear "what we do and where" statement on the homepage that an assistant can quote in one sentence, FAQ content that answers real customer questions. Nothing about the CMA proposal changes this list — it just puts a clock on it.

2. Audit your AI-bot blocking. In June we published a guide to blocking AI bots, because plenty of sites were blocking everything in a panic. If you're still running a blanket block, reconsider. Under the emerging UK framework, the assistants are obligated to attribute you. Blocking yourself out of that deal is leaving money on the table. Keep blocking the scrapers that give nothing back; let the assistants that must cite you in.

3. Start measuring assistant referrals now. Most analytics setups still lump ChatGPT, Perplexity, Copilot and Gemini referrals into "direct" or miss them entirely. Set up referrer tracking for the assistant domains before the choice screens land, so you have a baseline. You cannot manage a channel you cannot see, and the founders who spot the shift in their own data will react a year before the ones waiting for an industry report.

4. Own your entity, not just your keywords. Assistants don't rank pages; they answer about entities — businesses, products, people. Make sure your business is described consistently everywhere: same name, same services, same locations on your site, your Google Business Profile, directories, and social profiles. Inconsistent entity data is the #1 reason we see good businesses get skipped or misdescribed by assistants. It's unglamorous work. It matters enormously.

5. Don't chase the UK market specifically — chase the pattern. Unless you sell to Britain, the CMA's rules don't touch you directly. But the pattern — assistants as search defaults, mandatory attribution, yearly re-choice — will travel. The businesses that treat this as a UK curiosity and the businesses that treat it as a preview will be in very different positions in eighteen months.

One more position, since this is the kind of piece where fence-sitting is dishonest: we think the CMA is right. The default-search slot is too valuable to be permanently assigned, and yearly re-choice is a reasonable price for a company with a 90%+ share of the market it dominates. Google's products are excellent; they should win the re-prompt on merit. If they can't, that tells you everything.

FAQ

What did the CMA propose about AI assistants and search on 23 September 2026?

The UK's Competition and Markets Authority proposed that Google show Android and Chrome users a choice of search providers — including AI assistants like ChatGPT and Perplexity, if they meet technical and security criteria — when setting up a device or opening the browser, plus a once-a-year prompt to pick a default. The consultation runs until 9 October 2026, with a final decision expected by the end of the year.

Could ChatGPT become a default search engine?

Not quite yet — but the CMA's proposal is the closest any regulator has come to treating an AI assistant as a substitute for a search engine. If the rules are finalised, UK Android and Chrome users would be offered ChatGPT or Perplexity as a default option alongside Google and others, which would make AI assistants a mainstream search channel rather than a side activity.

What does the fair attribution requirement mean for website owners?

Any search provider on the choice screens must fairly attribute publisher content so users can see the source material. In practice this means AI assistants will have to credit and link back to the sites they draw on — good news for businesses whose content feeds AI answers, and part of a wider trend after publishers gained AI-result opt-outs earlier in 2026.

Does this affect businesses outside the UK?

Directly, no — the rules would apply to UK users only. Indirectly, they matter everywhere: regulators tend to copy each other, and the UK is the first to classify AI assistants as search substitutes on the default-search screen. If you sell in the US, UK or Pakistan, the smart move is to treat AI-assistant visibility as a standard channel now, not a novelty.

How is this different from the EU's Google choice screens?

Europe's Digital Markets Act already forced choice screens for search engines on Android. The UK proposal goes one step further by explicitly opening the screen to AI assistants, treating ChatGPT and Perplexity as alternatives to Google Search itself rather than separate products.

What should businesses do to prepare?

Three moves: first, make sure your site is already visible inside AI assistants — structured data, clear entity information and quotable facts, which is exactly the work we described in our SEO for AI search playbook. Second, stop blocking AI crawlers by default; blanket blocking was never free, and now it can lock you out of an emerging search channel. Third, start measuring referrals from ChatGPT, Perplexity and Gemini so you can see the shift as it happens.

When will we know the final decision?

The CMA is accepting views on the proposals until 9 October 2026 and expects to reach a final, binding decision by the end of the year. Implementation details and timelines would follow that decision.

The default is dead. Long live the re-choice.

The deepest change here isn't technical and it isn't even really about AI. It's about defaults. For twenty years, the most valuable sentence in consumer tech was "this is your default search engine." The CMA's proposal says that sentence now expires every twelve months.

Some founders will read this and file it under "UK regulatory news, not my problem." We think that's a mistake. The UK just drew the map everyone else will walk: assistants as search substitutes, mandatory attribution, defaults that have to be re-earned. The question for your business isn't whether this arrives in your market. It's whether you're visible inside the assistants when it does.

We build websites that rank on Google and get cited by the assistants that are trying to replace it. If your site isn't ready for the world where ChatGPT is somebody's default search engine, talk to us — and read how we approach the work on our services page. The consultation closes on 9 October. The shift it describes is already underway.

Sources

• Google search choice screen to add AI assistants, says CMA — HotMinute
• UK CMA Wants AI Assistants on Google's Choice Screens — AI Weekly
• Google faces UK push to put AI assistants on Android, Chrome choice screens — The Hindu BusinessLine (Reuters)
• CMA Strengthens Google Choice Screen Proposals — The Outpost AI
• Google Faces UK Plan to Open Android and Chrome Search to AI Rivals — PYMNTS