AI visibility tracking is about to get a rethink. New data from Similarweb and StatCounter shows the AI search market is consolidating fast around three platforms, but that does not automatically mean every other platform can be safely dropped from your reporting dashboard.
According to Similarweb’s May 2026 data, ChatGPT commands 53.9% of worldwide web visits among seven major AI assistants. Gemini follows at 27.9%, Claude at 9.2%, DeepSeek at 4.1%, Grok at 2.4%, and both Perplexity and Microsoft Copilot at 1.3% each. The concentration at the top is real. The question is what to do about it.
What triggered the debate about dropping Perplexity?
Ross Hudgens, CEO of Siege Media, made a direct call on LinkedIn, arguing that Perplexity’s market share is shrinking fast and that including it alongside ChatGPT, Gemini, and Claude gives marketers a distorted picture. His conclusion, as reported by Search Engine Journal: “Everyone should remove Perplexity from their LLM trackers today.”
The distortion concern is legitimate. If an AI visibility platform weights Perplexity equally to ChatGPT, an outlier result on Perplexity can move the aggregate score in a way that has no relationship to where real audiences are.
StatCounter’s worldwide AI chatbot referral share data (the share of clicks that each platform sends out to other websites) shows how quickly Perplexity has moved. In June 2026, Perplexity held 7.91% of AI chatbot referral share, essentially tied with Gemini at 7.94%. By August 2026, Perplexity had dropped to 4.31%, while Gemini rose to 10.9%. That shift in two months is hard to ignore.
Why referral share and visit share tell different stories
The two datasets measure different things and can point in opposite directions. Similarweb’s visit data captures how many people open each platform. StatCounter’s referral data captures how often each platform sends people to other websites. Perplexity’s referral share looked competitive against Gemini in June, even though Gemini’s visit share was more than 20 times larger.
The reason is structural. Perplexity’s interface is citation-heavy by design (it typically lists source links alongside answers). This inflates its referral share relative to its user base. Gemini and ChatGPT increasingly resolve queries inside the platform without requiring a click-out. High usage on those platforms does not always produce high referral share. For publishers and brands tracking clicks from AI platforms, referral data still matters, regardless of which platform sends those clicks.
How Gemini’s 1 billion users milestone changes the tier gap
In August 2026, Google announced that the Gemini app had surpassed 1 billion monthly users, making it the fastest-growing product in Google’s history. According to Google’s announcement, 63% of Gemini users now interact via voice, over 100 million active users are on iOS, and Gemini generates more than 150 million images every day. The app can also automate actions across 40 or more popular apps.
Scale of this kind changes the AI visibility calculation. A brand that appears in Gemini’s responses is now potentially reaching an audience comparable to a major social platform. Gemini is no longer a challenger to ChatGPT in user volume the way it was in 2025. It occupies a different tier from the rest of the field, and that gap widened sharply between June and August 2026 on referral share alone.
What Claude’s enterprise expansion signals about AI visibility
Claude’s story is playing out on a different track. Anthropic and PwC announced an expanded strategic partnership, with PwC rolling out Claude Code and Cowork to teams globally. The companies plan to train and certify 30,000 PwC professionals on Claude, and PwC is launching a new finance business unit built on Anthropic’s technology. According to Anthropic, Claude is already running in production across insurance underwriting, security operations, and software delivery, cutting delivery times by up to 70%. Insurance underwriting that previously took 10 weeks now takes 10 days.
Claude’s 9.2% visit share in Similarweb’s May 2026 data reflects a platform still building consumer scale. For B2B brands and professional services, though, Claude’s relevance as a visibility channel may outpace what those raw numbers suggest. Enterprise adoption concentrates usage among high-value, decision-making audiences.
Microsoft Copilot: another third-tier platform worth watching
Copilot sits at 1.3% alongside Perplexity in Similarweb’s May 2026 visit data. At the same time, Microsoft is landing large-scale enterprise deployments. According to Microsoft’s AI in Action news page, Westpac is rolling out AI to 35,000 employees and contractors, and the University of Manchester has become the first university to provide Microsoft 365 Copilot access to all 65,000 students and staff. Enterprise rollouts of this scale can shift platform usage data before consumer metrics reflect the change.
Is a three-tier market forming rather than a settled leaderboard?
Search Engine Journal’s analysis of the Hudgens argument raises a more useful framing: the data points toward a three-tier structure rather than a clear cut-off between “track this” and “ignore that.” ChatGPT sits alone at the top. Gemini and Claude form a competitive second tier. DeepSeek, Grok, Perplexity, and Copilot compete in a third tier with significantly smaller footprints.
Greg Jarboe, writing for Search Engine Journal, draws a historical parallel. At a search conference in March 2002, he recalls a proposal to simplify tracking to the five major search engines of the day: Yahoo, Excite, Lycos, AltaVista, and Ask Jeeves. Google was not on the list. It was gaining share rapidly, but it was not yet an obvious incumbent. The point is not that Perplexity will become the next Google. The point is that a market looks settled until it isn’t.
What this means for AI-search visibility
Here is the simplest way to think about it: your AI visibility tracker is like a poll. A poll that gives Perplexity the same weight as ChatGPT is like giving a city of 1,000 people the same electoral weight as a city of 40 million. The aggregate score becomes noise. But removing the smaller city entirely means you won’t notice if it starts growing quickly.
The more useful question is not “keep or drop Perplexity.” It’s whether your tracker weights platforms by their actual relevance to your audience. A B2B cybersecurity firm whose buyers use Perplexity for technical research has a different calculation than a consumer brand optimizing for mainstream assistant queries. The platform list should follow the audience, not the other way around.
One observation that tends to get skipped in the current debate: the referral gap and the visit gap may be diverging for structural reasons that won’t close quickly. As Gemini and ChatGPT push further into in-platform resolution (answering questions without requiring a click-out), their referral share will lag their visit share even as their user bases grow. That means referral-based AI visibility tracking and visit-based tracking will increasingly tell different stories for the same platform. Brands tracking which AI platforms drive clicks to their content need referral data; brands tracking brand mentions or citation rates need something else entirely.
Hingewise’s read: the case for dropping Perplexity entirely is weaker than the case for reweighting it accurately. A tracker that weights platforms by actual visit share, and reports referral share separately, gives a more honest picture of where AI visibility matters. The platforms worth watching most closely heading into Q4 2026 are those moving within the middle tier, where consolidation is still in motion and positions are not yet fixed.
Before you change your AI tracker setup
- Confirm which metric your current tool reports: referral share, visit share, or mention/citation rate. These measure different things and lead to different conclusions.
- Check your audience overlap with each platform before removing any from your dashboard. A platform small in aggregate may be large among your specific buyers.
- For B2B brands, track Claude-related visibility separately. Enterprise adoption means its relevance to professional audiences may exceed its consumer visit share.
- Run your aggregate AI visibility score with and without Perplexity to see if it materially changes your trend lines. If it doesn’t move the needle, downweighting is safer than removing.
- Ask your tracker provider how it weights platforms and whether weights are adjusted dynamically as market share data changes.
- Prioritize Gemini in your reporting: 1 billion monthly users and a referral share that climbed from 7.94% to 10.9% in two months (StatCounter, June to August 2026) make it the strongest challenger to ChatGPT across most markets.
- Set a monthly review cadence for platform weights. The StatCounter and Similarweb data is shifting fast enough that a one-time decision made in September 2026 may be outdated by November.
The more durable approach is building or choosing a tracker that can reweight platforms as data changes, rather than one that periodically restarts the debate over which platforms belong on the list at all.
Sources
- Search Engine Journal, Greg Jarboe, 2026
- Google Blog, Gemini 1 billion monthly users announcement, August 2026
- Anthropic, PwC expanded partnership announcement, 2026
- Microsoft News, AI in Action, 2026
Lam Nguyen · Hingewise
This article was drafted with AI assistance from the sources cited above and checked against them before publishing. The "What this means for AI-search visibility" section is Hingewise's editorial analysis, built on the frameworks and positions we use in our GEO work. How we work: Editorial policy.