Digit.fyi, reporting YouGov BrandIndex data, puts ChatGPT first at 30.7% in July, ahead of Alexa at 12.7%, Microsoft Copilot at 11.6%, Gemini at 11.3%, and Claude at 4.3%. In the headline table, ChatGPT’s result is rounded to 30%; later figures give the more precise 30.7%. Gemini’s 11.3% overall ranking and its 11.4% July trend figure are also both plausible, but they are evidently drawn from different aggregations or rounded reporting periods.
The result establishes a clear consumer-brand lead for OpenAI in Britain. It does not establish that nearly one in three UK adults uses ChatGPT most, or that ChatGPT commands 30.7% of all AI activity. Those are different questions, and the distinction changes what IT teams, product managers, and marketers can reasonably infer from the numbers.
YouGov’s metric measures a choice after consideration
YouGov’s Preference metric is a two-stage measure. Respondents first identify AI brands they would consider using, then name the one they most like or would prefer among that set. It is a useful measure of brand pull: when someone knows enough about several options to make a choice, which one gets selected?
It is not a usage share. A user may prefer ChatGPT yet spend a workday in Microsoft 365 Copilot because that is the tool their employer provides. Another may describe Gemini as their preferred assistant while relying on Google Search, Gmail, Android, and Pixel features that do not appear to them as a discrete Gemini session. Alexa complicates the category further: it is a long-established voice-assistant brand embedded in speakers and smart-home routines, not simply a general-purpose chatbot app competing on identical terms with ChatGPT or Claude.
That mixed category is the first limitation readers should keep in mind. The top five combines standalone AI services, enterprise-adjacent copilots, a phone-integrated assistant, and a smart-speaker platform. It reflects how the public experiences AI brands in 2026, but it should not be read as a clean model-quality league table.
YouGov’s public AI Ratings page illustrates why these measures cannot be swapped casually. Its second-quarter UK tracker lists ChatGPT with 87% “fame” and 46% “popularity,” while Gemini records 66% fame and 33% popularity. Those are broader public measures, calculated across adults and defined differently from the BrandIndex preference scores reported by Digit.fyi. A reader seeing 46% for ChatGPT in one YouGov product and 30.7% in another is not looking at contradictory polling; they are looking at different populations and questions.
The reporting also does not publish the sample size, confidence intervals, exact rolling windows, or the complete questionnaire for the March-to-July comparison. YouGov says its AI Index collects daily brand-health data from its panel and tracks metrics including consideration, usage, recommendation, trust, and perceptions. That makes rapid directional changes possible, but without the underlying fieldwork tables, the figures should be treated as a robust signal of relative movement rather than a precise census of British AI behaviour.
Gemini gained momentum, but advertising is not proven as the cause
Gemini’s move from 7.0% preference in March to 11.4% in July is the strongest increase in the reported set. In percentage-point terms, it gained 4.4 points; relative to its March base, that is roughly a 63% increase. For a brand that still trails ChatGPT substantially, the pace is the notable part.
Digit.fyi links Gemini’s rise with a period of heightened UK visibility, including Google Pixel advertising, the “Ask More of Your Phone” campaign, ODEON activity, and marketing involving footballer Trent Alexander-Arnold. Google’s UK Pixel materials do confirm that “Ask more of your phone” is current campaign language and place Gemini at the center of Pixel’s proposition.
But correlation is all the published record supports. YouGov did not attribute Gemini’s increase to a particular advert, device campaign, model release, or Google product integration. It would be premature to call this an advertising win, much less evidence that cinema promotions converted users from ChatGPT.
There are several other plausible explanations. Google expanded Gemini’s visibility through Android, Pixel, Search-related AI features, and its productivity products during the period. Consumers may also be responding to a changed product, not merely a changed message. And a preference survey can shift as people become familiar enough with a brand to include it in their consideration set; that is a different mechanism from users abandoning one assistant for another.
Still, the numbers show that Gemini’s gains have crossed the threshold from anecdotal momentum to detectable brand movement. Google’s advantage is distribution. It does not need every consumer to seek out a new AI destination if Gemini appears inside devices and services they already use.
ChatGPT’s five-point fall does not mean it lost the UK
ChatGPT fell from 35.8% in March to 30.7% in July, based on the figures cited by Digit.fyi. The article describes that as a five-point decline, but the arithmetic is 5.1 percentage points. It remains by far the largest movement downward in the data, ahead of Claude’s 4.4-point decline, Perplexity’s 3.7-point fall, and Copilot’s 2.6-point fall.
Yet the practical reading is erosion from a dominant position, not displacement. At 30.7%, ChatGPT’s preference score remains almost 19 points above Alexa, its nearest listed rival. Gemini’s rapid gain brought it to roughly a third of ChatGPT’s level, not into immediate contention for first place.
Independent usage-oriented data points in the same direction without matching the exact percentages. Statcounter’s UK AI chatbot traffic figures for June put ChatGPT at 68.12% and Gemini at 8.6%. That gap is far wider than YouGov’s preference spread because Statcounter measures web referral traffic rather than surveyed brand preference, but both datasets identify ChatGPT as the current leader and Gemini as a smaller challenger.
For Windows users, the more consequential issue is that ChatGPT’s brand lead has not translated automatically into a corresponding workplace lead for Microsoft. Microsoft retains the distribution advantage in Windows 11, Edge, Teams, Microsoft 365, and commercial licensing, but YouGov’s results suggest that consumer mindshare and daily work software remain separate battles. A Windows Copilot button or a Microsoft 365 entitlement does not necessarily make Copilot the assistant a user would pick first.
Copilot and Claude have an office-worker advantage
YouGov’s segmentation between knowledge workers and frontline workers is more useful for IT decision-makers than the overall ranking. Copilot’s preference score was 15.9% among knowledge workers, compared with 8.6% among frontline workers. Claude showed a similar split: 7.5% among knowledge workers and 2.1% among frontline workers.
That pattern fits the products’ real-world positioning. Copilot is closely tied to document creation, meetings, email, code, data analysis, and Microsoft 365 workflows. Claude has developed a reputation among professional users for long-form analysis, writing, coding, and document handling. Neither needs to win the broad consumer market to be strategically significant in higher-value office workflows.
YouGov’s own February research on AI at work found that 32% of workers in Britain said they used AI in their current job, with Copilot and ChatGPT among the prominent tools. But the new preference figures cannot prove which assistant workers use at work. YouGov specifically cautions that the knowledge-worker and frontline-worker results cover overall preference and can reflect personal as well as professional usage.
Alexa runs in the opposite direction, with 14.4% preference among frontline workers and 6.4% among knowledge workers. Gemini also scores more strongly among frontline workers, at 14.7% against 12.6% for knowledge workers. Those results reinforce the danger in treating this as a straightforward enterprise-AI ranking: accessibility through phones, voice interfaces, and consumer devices may matter more than office-suite integration in the broader workforce.
Accuracy perceptions show the value of awareness
ChatGPT also led YouGov’s reported Accuracy metric with a net score of 7.3, followed by Gemini at 5.6, Copilot at 3.6, Alexa at 3.4, and Claude at 1.9. On Personal Impact, the gap widened: ChatGPT scored 19.0, compared with Copilot’s 10.1, Gemini’s 8.1, Alexa’s 6.1, and Claude’s 2.9.
The Claude result reveals an important survey effect. Its overall Accuracy score of 1.9 rises to 5.2 among respondents who are already aware of Claude, according to the Digit.fyi account of the YouGov findings. Claude’s low all-user score therefore says at least as much about reach as it does about user judgment. A tool cannot accumulate broad public trust among people who do not know it exists.
That is where the consumer AI contest is likely to become harder for specialists. ChatGPT began with a huge awareness advantage; YouGov’s public Ratings data still shows it well ahead of Gemini and Claude on fame. Gemini is now using Google’s distribution to narrow the consideration gap. Copilot can use the Microsoft estate, particularly where employers turn on the paid service and provide training. Claude and Perplexity must convert perceived usefulness among knowledgeable users into wider recognition without the same operating-system, search-engine, phone, or smart-home foothold.
For now, the UK results point to a simple operational conclusion: organisations should not select an AI strategy based on a consumer preference league table, but they should not ignore it either. ChatGPT owns the public default; Gemini is gaining rapidly through visibility and distribution; Copilot retains its strongest case where Microsoft work is already being done. The next procurement decision will be shaped less by which brand wins a national preference poll than by whether employees can use the approved tool well enough to avoid bringing their preferred one in through the side door.