The best-known AI tool is not automatically the one people prefer, trust most or find most useful. UK consumer data from 2026 shows a market with a clear leader, fast-moving challengers and meaningful differences between the people using these tools.
For product and technology leaders, the useful question is not simply “Which AI brand is first?” It is “What does first place actually measure, and what should we measure alongside it?” This article looks at preference, perceived accuracy, personal impact and audience context as separate signals of product strength.
ChatGPT leads the UK preference table
ChatGPT is the clear leader among UK AI users, with 30.0% saying it is the AI assistant they prefer most. Alexa follows at 12.7%, with Copilot at 11.6%, Gemini at 11.3% and Claude at 4.3%.
The top ten preference table is:
| Rank | Brand | Preference |
|---|---|---|
| 1 | ChatGPT | 30.0% |
| 2 | Alexa | 12.7% |
| 3 | Copilot | 11.6% |
| 4 | Gemini | 11.3% |
| 5 | Claude | 4.3% |
| 6–10 | Siri, Grok, Meta AI, Apple Intelligence, Perplexity AI | Top ten |
Preference is narrower than awareness or consideration. People are first asked which tools they might consider using, then which of those options they like or prefer most. It therefore gives a better view of the first choice among plausible alternatives than a simple awareness ranking does.
ChatGPT’s lead is substantial, but the rest of the market is much more tightly grouped. That creates a strong advantage for the leader while leaving plenty of room for challengers to move as familiarity, distribution and product experience change.
A leading position can still be fragile
The comparison between March and July 2026 suggests that the category is still settling. Gemini records the largest rise in preference, moving from 7.0% to 11.4% — a gain of 4.4 percentage points. DeepSeek gains 0.5 points and Meta AI gains 0.1 points.
Over the same comparison, ChatGPT remains the UK leader but falls from 35.8% to 30.7%, a 5.0-point decline. Claude falls by 4.4 points, Perplexity AI by 3.7 points and Copilot by 2.6 points.
These movements should not be treated as proof that one campaign or feature caused a change. They do show that visibility and access matter. Gemini’s gains coincide with a period of stronger exposure across consumer devices, search and marketing activity. The practical lesson is that a capable product still needs repeated, relevant opportunities for people to discover and try it.
The small difference between ChatGPT’s 30.0% in the main table and 30.7% in the July comparison is a reminder to check the measurement window before interpreting a trend. A ranking covering a longer period is not directly interchangeable with a single-month result.
The shortlist changes with the working context
People in different roles do not necessarily choose the same AI tools. ChatGPT is relatively consistent across the two worker groups, while several other brands show a clear split:
| Brand | Knowledge workers | Frontline workers |
|---|---|---|
| Copilot | 15.9% | 8.6% |
| Claude | 7.5% | 2.1% |
| Alexa | 6.4% | 14.4% |
| Gemini | 12.6% | 14.7% |
Copilot and Claude perform more strongly among knowledge workers, which is consistent with their association with office productivity, writing and other screen-based tasks. Alexa is stronger among frontline workers, where voice interaction can be useful while driving, carrying equipment, preparing food or working away from a desk.
Gemini is also slightly stronger among frontline workers. The data does not establish why, but everyday familiarity through mobile devices, search and other consumer touchpoints may be part of the explanation.
For product teams, the message is simple: an “average user” score can hide important differences in the conditions in which a product is used. Role, environment, device and task should be part of the segmentation plan when evaluating an AI feature.
Preference, accuracy and impact answer different questions
A product can be the preferred choice without being seen as the most accurate. It can also improve someone’s work without being the tool they would choose first. A useful AI scorecard therefore separates three questions:
- Preference: Which tool would the user choose over the available alternatives?
- Accuracy: Does the user believe the tool provides reliable information?
- Personal impact: Does the tool have a positive effect on how the user works, thinks or manages everyday tasks?
On perceived accuracy, ChatGPT leads with a net score of 7.3. Gemini follows at 5.6, ahead of Copilot at 3.6, Alexa at 3.4 and Claude at 1.9. The net score represents positive perceptions minus negative perceptions.
The gap is wider on personal impact. ChatGPT scores 19.0, followed by Copilot at 10.1, Gemini at 8.1, Alexa at 6.1 and Claude at 2.9.
That difference matters. Users may prefer a familiar tool while remaining cautious about the reliability of its answers. Another tool may not be the first choice overall but still create clear value in a particular workflow. Treating preference as a proxy for every other outcome leads to weak product decisions.
Awareness also changes the interpretation. Claude has lower overall awareness than the other brands in the top five, but its perceived accuracy is stronger when the analysis is limited to people who already know the brand. That suggests some of the gap may be about reach and familiarity rather than product experience alone.
What product leaders should take from the data
Leaders need to protect trust as well as share
A large preference lead creates valuable momentum, but it also raises the standard for consistency. Leaders should keep investing in response quality, transparency, data controls and predictable user experiences rather than assuming that habit will protect the position indefinitely.
Challengers need useful visibility, not just more visibility
Gemini’s movement illustrates the value of distribution and repeated consumer contact. But awareness only creates an opportunity to try a product. Long-term preference depends on proving that the product performs a meaningful task better, more safely or with less friction.
Segmentation is more useful than a single average
Office staff, operational teams, managers and students may use the same AI brand for entirely different reasons. Track preference, accuracy and impact by audience so that product investment follows a real need rather than a blended market average.
AI investment should connect to a real workflow
“Which model is most popular?” is a reasonable starting question, but it is not a delivery plan. A stronger assessment asks:
- Which user group and task are we improving?
- What is the cost of an inaccurate answer, and how will we control it?
- Is the product saving time or simply adding another tool to learn?
- How will preference, usage, reliability and business outcomes be measured together?
These questions move an AI decision away from brand popularity and towards measurable product value. DuniaOps’ AI software development service helps teams work through that path from discovery and prototyping to dependable production delivery.
Methodology and limitations
The main indicators discussed here cover UK AI users between March and July 2026. The preference, accuracy and personal impact rankings use minimum bases above 6,740. The worker comparison uses minimum bases above 2,120 for knowledge workers and 650 for frontline workers.
The movement analysis compares March and July. Alexa and Siri were not included in the category tracking in March, so they are excluded from that comparison. The published ranking focuses on consumer-facing, general-purpose AI assistants rather than specialist image-generation tools, research companies or model families.
These figures are a snapshot of market perception, not a product selection or investment recommendation. Models, pricing, features and user expectations change quickly, so teams should repeat the same measures over time and review them alongside actual usage and business outcomes.
The practical conclusion
ChatGPT is the clear UK preference leader in 2026. Gemini is gaining momentum, Copilot and Claude have stronger associations among knowledge workers, and Alexa stands out in more practical, voice-led contexts.
The broader lesson is more useful than the leaderboard: preference, perceived accuracy and personal impact are not interchangeable. Sustainable advantage in AI comes from combining reach, a credible product experience, trustworthy answers and a clear improvement to the user’s work.
If you are deciding where AI can create value in your organisation, or how to improve an existing system safely, speak to our software consultancy team.
