When a buyer asks AI what to buy, does your brand come up?
onChatGPT
JalsaX AI asks the questions your buyers ask, on the assistants they use, and reports how often the answers name you. Every figure comes with the sample behind it and how far it could be off.
- Every figure with its range and sample count
- Five assistants, asked from India
- Analytics and advisory, no ranking promises
The answer is the last thing they see.
Search reporting shows where you rank on a results page. It does not show whether an assistant named you, or what it said.
- 43%
- of Indian online searchers use an AI assistant to compare options before deciding.YouGov, n=1,004 urban adults, fielded April to May 2026
- 28%
- stop searching entirely once the assistant answers. That answer is the last thing they see before deciding.YouGov, the same study
- 1 in 100
- runs of the same prompt returned an identical brand list. Asking once tells you almost nothing.SparkToro, 2,961 prompts, 2026, with no India-specific sample reported.
One real month, with the uncertainty drawn.
A skincare brand in India, three waves in August 2026, 90 buying questions. Every figure is a band, because one answer is one moment.
Three waves, the same 90 questions
Named in about four answers in ten, and the waves agree. No pair differs by more than the sample can tell apart.
- Wave 141.1% 30.9% to 51.3% · 37/90
- Wave 250.0% 39.7% to 60.3% · 45/90
- Wave 344.4% 34.2% to 54.7% · 40/90
Question by question, 90 questions
- Named in all three waves24 of 90
- Named in two15 of 90
- Named in one20 of 90
- Named in none31 of 90
35 of 90 questions flipped at least once across the three waves with nothing being done to them. The total held; the single answer did not. That is why this is monthly, and never a daily score.
Six categories, 45 answers each
Strong on sunscreen, largely absent from body lotion. The middle four are one group; this sample cannot order them.
- Sunscreen66.7% 52.9% to 80.4% · 30/45
- Face wash60.0% 45.7% to 74.3% · 27/45
- Actives60.0% 45.7% to 74.3% · 27/45
- Vitamin C serum51.1% 36.5% to 65.7% · 23/45
- Moisturiser20.0% 8.3% to 31.7% · 9/45
- Body lotion13.3% 3.4% to 23.3% · 6/45
Shaded: differences within 27 points of the weakest category are not resolved at this sample size. Only the two extremes are separated, and the two weakest are absent together, which points at one gap rather than two.
14 brands the assistant named
One tier, not a league table. No adjacent gap is larger than the 11 points this month can resolve.
The measured brand is the third-highest share, at 13.4%. Positions two, three and four could be in any order and this data could not tell the difference. One further declared brand was named in no answer at all.
What the assistant read
The pages read when you are missing are the pages to be on. Cited when the brand was absent, against when it was named.
- A competitor's own site17.1% vs 0.0%
- A shopping site18.0% vs 12.1%
- A second competitor's site9.0% vs 0.0%
- A health site9.0% vs 5.2%
- A fashion magazine7.2% vs 0.0%
- A beauty magazine5.4% vs 0.0%
brand absent, 111 answersbrand named, 58 answers
Skincare benchmark, waves 1 to 3, August 2026. Prepared by JalsaX, not affiliated with any brand measured. Two assistants answered this month; the other three are reported as unavailable, never as a zero. Percentages are of answers; every band is a 95% interval, and the counts beside it are the answers behind the figure.
What arrives each month
A report you can check, not a score you have to trust. The figures below are the benchmark’s own, not a picture of a screen.
One report, one brand, one month
Every figure with its sample and its range, and a plain line on what could not be measured.
Coverage, August 2026
Which assistants answered the benchmark's questions this month.
Every report contains
- Mention rate, with its 95% interval and the sample count behind it
- How each assistant described you, and the sentiment of it
- The sources the answers drew on, and which ones appear only when you are missing
- Where you sat in the list when you were named
- Per assistant, whether it answered at all that month
- Category by category, where the visibility is and where it is not
Monthly, and the report says which month it describes. Each assistant is reported on its own; one that did not answer is reported as unavailable, never as a zero.
How a month is read
Two people asking the same assistant the same question on the same day can see different brands.
- 01
The questions your buyers ask
A fixed set of buying questions, written for how people in India ask, and held the same month to month, so a change in the number is a change in the answers.
- 02
Asked from India, again and again
Every question goes to every assistant repeatedly, from India, because one answer is one moment. The repeats are what turn a snapshot into a measurement.
- 03
Read by a fixed rule
Whether an answer named you is decided by a fixed rule, the same way every time, never by a model's opinion. An assistant that did not answer is recorded as unavailable, never as a zero.
- 04
Reported with its range
Mention rate with its 95% interval and the sample count, how you were described, what the answers cited, where you sat in the list, and for each assistant whether it answered at all.
Five assistants, each reported on its own
- ChatGPT
- Gemini
- Perplexity
- Claude
- Google AI Overviews
An assistant that does not answer in a given month is reported as unavailable, with the count of attempts. It is never filled in from somewhere else, and it is never a zero for your brand.
What JalsaX AI will not do
The firm that measures your visibility should not also be the firm paid to move it. These three hold for every brand and every month.
It does not touch your content.
We do not rewrite pages, execute changes, or submit anything to an AI platform on your behalf. We measure and report, and every decision stays with your team.
It does not guarantee a ranking or a recommendation.
Nobody can promise an assistant will name you, and a report that implies otherwise is selling something else. JalsaX AI is analytics and advisory.
It does not show you a daily number.
Assistants answer differently run to run. A daily figure would move every day and none of the movement would mean anything, so the report is monthly and says which month it describes.
Questions brands ask us
Who is JalsaX AI for?
Brands that customers ask an assistant about before they buy. The first cohort is skincare and beauty in India, because that is the category where the questions are most specific and the claims most contested.
Which assistants do you measure?
ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews. Each is reported separately, and an assistant that did not answer in a given month is reported as unavailable, never as a zero for your brand.
Why does every number come with a range?
Because assistants answer differently run to run. Every figure arrives with the number of answers behind it and a 95% confidence interval, including the months where that makes the number look less impressive. A number without one invites you to act on noise.
Why not a daily score?
A daily figure moves constantly and none of the movement means anything at this kind of sample. Monthly waves on the same questions are what make a real change detectable.
Will JalsaX AI improve how often we are named?
No. It measures and tells you what it found, and the decisions and the work stay with your team. Anyone promising an assistant will name you more often is guessing, and no ranking guarantee is on offer here.
What does the first month include?
The first month is free: the full report for one brand on the assistants that answered, with every figure carrying its range and sample count. What follows is agreed with you in writing before anything is invoiced.
Is the benchmark on this page a client?
No. It is a public benchmark JalsaX ran on a skincare brand in India in August 2026, shown here without the brand's name and with no affiliation. The figures are shown exactly as measured.
How is this different from an SEO report?
Search reporting shows where you rank on a results page. It does not show whether an assistant named you at all, how it described you, or what it read before deciding.
Assistants already answer questions about your category. Start with one month of the answers.
Your search reporting shows where you rank on a page of results. It does not show whether an assistant named you at all.
Start with a free monthSkincare and beauty first · monthly, not a daily score · the first month is free
Start with a free month.
Tell us the brand and the site, and we come back with the questions we would ask and what the first month would measure. The first month is free. Or email info@jalsax.com.
- 5 assistants, asked from India, each reported on its own
- Every figure with its 95% interval and the sample behind it
- Reply within 24 hours, on email or WhatsApp