According to the Celebrity Brand AI Demand Index from 5W and Talent Resources, Rare Beauty, founded by Selena Gomez, drew about 33,600 US ChatGPT prompts a month from August to October 2026, and Fenty Beauty, founded by Rihanna, about 24,400. Tatcha, with no celebrity founder, rose fastest, from about 14,100 prompts in August to about 70,700 in October. Talent teams should check AI demand before signing a beauty deal, not after.
Key takeaways:
- Rare Beauty averaged about 33,600 prompts a month and Fenty Beauty about 24,400, more than four of the five brands without a celebrity founder.
- Tatcha rose from about 14,100 prompts in August to about 70,700 in October.
- Prompt volume measures demand, not whether AI recommends a brand or whether its answer is accurate.
- Talent teams can add AI demand to a deal review with five questions, listed below.
What do people ask ChatGPT about celebrity beauty brands?
People ask ChatGPT about celebrity beauty brands tens of thousands of times a month. 5W and Talent Resources counted how often people mention seven beauty brands when they prompt ChatGPT in the United States, month by month, from August to October 2026. The research gives talent teams a number they did not have before.
Rare Beauty and Fenty Beauty each drew more prompts than four of the five brands in the study that have no celebrity founder, and only Tatcha drew more than either of them. The table shows the monthly average and the change from August to October for all seven brands.
US ChatGPT prompts that mention each brand, monthly average and change, August to October 2026. Source: 5W's analysis using Profound's tools.| Brand | Celebrity-founded | Monthly average | Change, August to October |
|---|
| Tatcha | No | 41,981 | +401% |
| Rare Beauty | Yes, Selena Gomez | 33,568 | +5% |
| Fenty Beauty | Yes, Rihanna | 24,397 | +1% |
| Sol de Janeiro | No | 23,555 | -58% |
| Laneige | No | 14,624 | -63% |
| Glow Recipe | No | 13,935 | -3% |
| Drunk Elephant | No | 13,583 | -14% |
I find the celebrity numbers encouraging. Rare Beauty drew 2.3 times the prompts of Laneige, and Fenty Beauty drew 1.7 times. A famous founder gives a beauty brand a name that people type into an AI tool without any prompting from an advertiser, and that is a real asset. For a talent team, those two numbers are a useful starting point for a first conversation with a brand about how it plans to turn recognition into answers.
Why did Tatcha's demand rise five-fold?
Tatcha's prompts rose from 14,117 in August to 41,120 in September and 70,707 in October, which is five times higher and twice Rare Beauty's October volume of about 35,400. 5W did not test the cause. Holiday gift shopping is one possible explanation, and the data cannot confirm it.
As a talent representative, I take two things from the pattern. Demand for a brand can move fast, and it can move for brands that have no celebrity attached. Over the same two months Laneige fell 63% and Sol de Janeiro fell 58%, while Rare Beauty stayed between about 31,400 and 35,400 prompts in every month and Fenty Beauty ended October within 1% of where it began.
The two celebrity-founded brands held steadier than most of their rivals. I would not build a deal on one month of data in either direction, but I would want to see the direction over the last three months before I recommended a partnership to a client.
What do the earlier Talent Resources and 5W studies add?
The demand study builds on earlier research from Talent Resources and 5W. Everything-PR published coverage of the Celebrity-Brand Fit Index when it launched. The Celebrity-Brand Fit Index ranks eight consumer sectors on five variables: consumer receptivity, verified return on investment, category fit, risk exposure and whitespace. It scored beauty 7.8 out of 10, second of eight sectors behind spirits and beverage, so beauty is a sector where celebrity involvement can pay off.
The Celebrity Endorsement Index looked at what AI engines say rather than what people ask. It reported that 18% of the celebrity-brand partnerships four AI engines recommended were fabricated, and that 64% of 212 brand marketers surveyed in May 2026 begin endorsement scoping inside an AI engine.
Put the pieces together and the picture is practical. Marketers start in AI, people ask AI about celebrity brands by name, and the answers are not always right. A good deal review has to cover all three.
What should talent teams ask before signing a beauty deal?
Five questions follow from the data. They are my suggestions, and the study did not test whether asking them improves a deal. I covered the wider checklist in five questions before greenlighting a celebrity deal.
- How many people ask AI about this brand by name each month, and which way is that moving? Tatcha rose five-fold and Laneige fell 63% in two months, so direction matters as much as level.
- What does AI say about the partnership? Ask ChatGPT, Gemini, Perplexity and Claude, then check every deal they name against press releases and filings.
- Which questions do people ask about the brand? Someone who types a brand name into ChatGPT is usually asking about a product, a shade or where to buy, and the brand should have pages that answer those questions directly.
- Who else writes about the brand? AI engines are more likely to cite sources that many independent publications reference, so reviews and retailer pages count alongside the brand's own site.
- How will we measure the deal? I would add prompt volume and citations to awareness, engagement and sales, and review them monthly.
What can prompt volume not tell a talent team?
Prompt volume counts how often people mention a brand name when they prompt ChatGPT. It does not show whether ChatGPT recommends the brand, what it says about it or whether people buy. The estimates are modeled, and a month backed by two or fewer distinct users is not reported.
The sample is also small. Only seven beauty brands returned a reportable estimate in all three months, two of them celebrity-founded, and 5W chose the brands. Rare Beauty and Fenty Beauty show what two celebrity-founded brands did over three months, and I would be careful about generalizing to every celebrity brand. A celebrity-founded brand here means one founded or co-founded by a public figure, regardless of later ownership, so the label does not mean the celebrity owns the brand today.
What should celebrity founders and their teams do next?
I draw two conclusions, and both are optimistic. First, a celebrity's name opens the door in AI, as the Rare Beauty and Fenty Beauty numbers show, so fame remains an asset. Second, the brands that keep the door open will be the ones that know the questions people ask and answer them.
Celebrity partnerships have traditionally been measured through awareness, engagement and sales. AI adds another measure: whether a brand appears when consumers are actively looking for recommendations. That is a new opportunity for talent, founders and marketers to build lasting brand value beyond the celebrity's name. I would encourage talent teams, and the celebrity marketing agencies that serve them, to add AI demand to the way they read a deal, including deals that are still in negotiation, since the data is available now.
Founders can do the same for their own brands. Check what people ask AI about the brand, answer those questions on pages you control, and look again every month through the holidays, because Tatcha's rise shows how quickly the numbers can change. Black Friday falls on November 27, so the next few weeks are a good time to start, before the holiday sale season reaches its peak.
The full tables, month-by-month data and methodology are in the Celebrity Brand AI Demand Index on 5W's research page. Talent Resources is the research partner on the series, and my Everything-PR profile lists the earlier studies.
