For agencies that grow D2C brands
How do you track a brand's AI search visibility in ChatGPT and Perplexity?
AI search visibility is how often answer engines such as ChatGPT and Perplexity name a brand, or cite its pages, when shoppers ask the questions that lead to a purchase. Track it as a rate across a fixed set of buying prompts, not as one screenshot, and note which rivals get named instead.
Named in 1 of 233 tracked AI answers, while its pages drew 344,000 impressions in Google's AI results · Moolwanai-dm · Sep 2026 · Search Console · Generative AI features · 25 Jun–23 Sep 2026 · real data
What you'll see in 10 minutes
- Your first brand's report, built from its domain alone.
- Its AI assistant visibility, read in the first starter research run.
- Its search visibility and brand mentions from the same runs, side by side.
What does AI search visibility actually measure?
Thursday's brand review. The founder asks, "Are we showing up in ChatGPT?" Someone on your team types one question, gets one answer and shares a screenshot. Next week the answer is different, and the brand starts wondering what else you are guessing.
Four numbers answer the question honestly. Each one tells you something different, and a brand can do well on one while failing another. These are Moolwan's real figures.
| Measure | What it tells you | Moolwan | Source |
|---|---|---|---|
| Mentions | How often the answer names the brand | 1 of 233 | ai-dm · tracked AI assistant answers · Sep 2026 |
| Citations | How often the brand's pages are linked as a source | 115 | ai-dm · citations across moolwan.com · Sep 2026 |
| Google's AI results | How often the brand's pages are shown in Google's generative AI features | 344k | Search Console · Generative AI features · 25 Jun–23 Sep 2026 |
| Visits from AI | How many people arrive on the site from an AI assistant | 368 | GA4 · First user source chatgpt.com · 29 Jun–26 Sep 2026 |
Cited is not the same as recommended. Moolwan's pages were cited 115 times, yet the brand was named in just 1 of 233 answers. The engines trusted its guides as sources without suggesting it as a place to buy. That gap is invisible in a single screenshot, and it is where the work is.
Which prompts should you track for a brand?
Track the questions that come right before a purchase. "What is a cushion cover?" rarely sells anything; "best handloom cushion covers under ₹1,500" does. Five kinds of prompt cover most buying moments.
Category prompts
"What are the best brands for this in India?" The broadest test of whether the brand is in the conversation at all.
Budget prompts
The category plus a price ceiling, which is how most D2C shoppers actually ask.
Comparison prompts
The brand against its main rival, and "alternatives to" the rival. Here you learn who the engine prefers.
Brand check prompts
"Is this brand any good?" What the engine says when a shopper already knows the name.
Use and gift prompts
The occasion or problem behind the purchase, such as a gift or a room to furnish.
Build a brand's buying prompts, then score them
Fill in the brand and get eight buying prompts. Run each in ChatGPT or Perplexity, mark what the answer named, and see the brand's mention rate against its rival.
- What are the best handloom cushion covers brands in India?
- Best handloom cushion covers under ₹1,500 in India
- Which handloom cushion covers would you recommend as a gift?
- Is Kora Loom Home a good brand for handloom cushion covers?
- Kora Loom Home or Weavecraft: which is better for handloom cushion covers?
- Good alternatives to Weavecraft for handloom cushion covers
- Where can I buy handloom cushion covers online in India?
- What should I look for when buying handloom cushion covers?
Run each prompt and mark the result to see where the brand stands.
Free, no login. The starting brand and rival are an illustrative example. One run per prompt is a snapshot; ai-dm tracks the same prompts over time.
Why does a brand show up one day and vanish the next?
AI answers are not fixed rankings. The same prompt can name different brands from one run to the next, change with small differences in wording, and shift when the sources the engine reads change. A single screenshot proves only that one answer happened once. Five habits keep the reporting honest.
Track the same prompts every week
Changing the questions changes the answer. A fixed set of buying prompts is the only fair comparison over time.
Report rates, not single answers
"Named in 6 of 20 answers" survives a bad day. One screenshot does not.
Keep mentions and citations apart
A cited page shows the engine trusts the brand's content; a mention shows it recommends the brand. They move separately.
Note which rivals are named, and which pages are cited
The rival's cited page is usually the clearest brief for what the brand should publish next.
Change one thing at a time, then watch the rate
Publish one page that answers one buying question, and give the engines a few weeks before judging it.
(login required) · Walk into the next review with a trend, not a screenshot.
How does ai-dm track AI search visibility?
The same buying prompts, tracked for every brand. ai-dm tracks the brand's answers in ChatGPT and Perplexity, counts how often the brand is named, which of its pages are cited and which rivals are named instead. Those rivals are usually the ones to watch in competitor ad tracking too. With Search Console and GA4 connected, it adds the brand's impressions in Google's AI results and its visits from AI assistants, and shows each brand's rate in one dashboard for every brand.
cites: weavecraft.in · a marketplace listing
cites: koraloom.in/guides/cushion-cover-sizes
named in 1 of 2 engines · rivals named in 2 of 2 · brand pages cited 1
In this illustrative example, ChatGPT answers "best handloom cushion covers under ₹1,500 in India" by naming two rivals and not the brand. Perplexity names the brand third and cites its guide to cushion cover sizes. Across both engines the brand is named in 1 of 2 answers and its rivals in 2 of 2.
From a gap to a page that closes it. When the engines name a rival and cite its guide, the fix is usually a page that answers that buying question better than anyone else. Moolwan's top guide, on TV placement by Vastu, drew 8,479 impressions in Google's AI results (Search Console · Generative AI features). When a gap shows up, ai-dm drafts articles and schema that answer the brand's buying questions, for your team to approve, and a daily change alert tells you when the brand's mention rate moves.
We show the gap on our own brand. Moolwan is named in just 1 of 233 tracked AI assistant answers (ai-dm · Sep 2026), even with strong Google AI results and 368 shoppers arriving from ChatGPT (GA4 · 29 Jun–26 Sep 2026). It stays on our own list of things that are not working yet. You can watch ai-dm run Moolwan, gaps included.
Related pages
Questions agencies ask about AI search visibility
How do you track a brand's visibility in ChatGPT and Perplexity?
Pick a fixed set of buying prompts for the brand, run them in each engine, and record whether the answer names the brand, cites its pages or names a rival. Repeat on a schedule and report the rate, not one answer. ai-dm runs this for every brand you add; the prompt builder on this page lets you try it by hand first.
What is the difference between being mentioned and being cited?
A mention is the answer naming the brand, often as a recommendation. A citation is the engine linking one of the brand's pages as a source. They move separately: Moolwan's pages were cited 115 times while the brand was named in just 1 of 233 tracked answers (ai-dm, Sep 2026).
What is AI share of voice?
AI share of voice is the brand's share of the mentions across a fixed set of prompts, compared with the rivals named in the same answers. If the brand is named 4 times and its main rival 12 times, the brand's share against that rival is 25%. Score a brand's own prompts with the builder on this page.
Can ai-dm show how a brand appears in Google's AI results?
Yes, once the brand's Search Console is connected. ai-dm reads the Generative AI features report, which counts how often the brand's pages are shown in Google's AI results. For Moolwan that was 344,000 impressions between 25 June and 23 September 2026, led by a single guide with 8,479.
Can anyone guarantee a brand will be recommended by ChatGPT?
No. No one controls what an AI engine says, and an agency that promises a placement is guessing. What you can do is track the rate honestly, find the pages the engines cite for rivals, and publish better answers to the same questions. ai-dm shows the gap; your team decides what to publish.
How often should we report AI visibility to a brand?
Track weekly and report the monthly trend: mention rate, citations, rivals named and the pages behind them. Monthly is long enough to see a real change and short enough to keep the brand engaged. The same numbers can go into the white-label reports your brands get under your agency's name.
No plan to pick. Billed in USD and INR. A brand's data is deletable on request.