Category Overview: what your market asks AI, every day
Category Overview is Intura's daily research on one product category in one market. It shows the real questions people ask about the category, how often and how widely they're asked, which brands and sites AI answers name and cite, the keywords in those answers, and the customer segments behind the questions. You follow a category, such as Beauty & Personal Care in Indonesia, and the research refreshes every day.
Nine categories, three markets: Indonesia, Malaysia and Vietnam. Below is the daily research cycle, step by step.
Free plan available · No credit card needed ·
How does Category Overview research a category?
Two daily processes run for every followed category and market: one finds and tracks the questions people ask, the other reads what AI answers say about them.
- Step 1
Follow a category in a market
Pick a market (Indonesia, Malaysia or Vietnam) and one of the nine categories listed below. A workspace follows up to three category-and-market combinations, for example Fashion & Style in Indonesia and Food & Beverage in Malaysia. Research for a newly followed category starts the same day; after that it runs every day.
Each market brings its own languages (Indonesian, Malay or Vietnamese, plus English and mixed-language posts), its own communities and its own regulators, so a skincare question in Kuala Lumpur isn't researched the same way as one in Jakarta.
- Beauty & Personal Care
- Food & Beverage
- Travel & Hospitality
- Fashion & Style
- Tech & Gadgets
- Parenting & Family
- Digital Products
- Fitness & Health
- Finance & Investment
- Step 2
Harvest real questions through three lenses
Every day, for every followed category and market, Intura runs grounded web research through the three lenses below, once for each of four funnel stages: awareness, consideration, decision and post-purchase. That is 12 research passes per category per day.
Each pass may only report questions it found real people asking in pages it actually retrieved, and it is told never to invent one. Questions must have been posted within the last six months and are kept in the language they were asked in.
- Communities and comments: forums, Reddit, Quora, Facebook groups, comment sections on TikTok, YouTube and Instagram, and marketplace Q&A. Local examples include Kaskus in Indonesia, Lowyat in Malaysia, and Voz and Tinhte in Vietnam.
- Search: "People also ask" boxes, autocomplete-style queries, related searches and the questions articles are written to answer.
- Complaints and reviews: the problems and doubts people describe right before or right after they buy.
- Step 3
Label each question and group by meaning
Each harvested question is labelled with an intent (from 12 intent families), a funnel stage, a topic, the pain point behind it, its keywords and any brands it mentions.
Questions that mean the same thing are grouped, so a question and its paraphrases become one question with several phrasings, shown as "other ways people ask it". Pain points are grouped the same way, so the same frustration worded ten ways counts once.
- Step 4
Keep only what independent passes agree on
One research pass can be wrong. A grouped question only moves forward if at least two separate passes found it. This consensus gate is the main filter against questions nobody actually asks, and it caps each day's intake at the 40 strongest question groups.
- Step 5
Verify each question on the web
Questions that pass the gate are checked again on the web, a few at a time. Verification records the demand side (how widely the question is asked on a 1–5 scale, where it was found and when it was last seen) and the answer side (how complete and consistent existing answers are, which brands they name and in what formats). Up to four evidence pages are kept, and only sources the research actually retrieved count.
Questions that verification rejects are dropped. Ones it can't confirm are kept but marked as weaker; confirmed ones show a "Verified on the web" mark.
- Step 6
Score demand, opportunity and trend
Each question gets a demand index from 0 to 1. It is a proxy built from how widely and how consistently the question was found, not search volume from a keyword tool.
Opportunity (0–100) weighs demand (35%), how poorly current answers cover the question (30%), how valuable the intent is (20%) and how open the answers are to new brands (15%). Questions are then ranked into deciles and tiers within their category.
- Step 7
Track every question over time
Each day's questions are matched to earlier ones by meaning. A question keeps its first-seen date, gets a new last-seen date every time it reappears, and leaves the active view after 30 days unseen. If it comes back, its history comes back with it.
Demand is compared over up to 30 days to mark each question new, rising, stable or declining, and the page counts how many active questions are rising, steady or falling.
- Step 8
Read what AI answers say about the category
The second daily process powers the Keywords, Brands and Segments views. Intura writes about 100 questions per category and market, spread across the four funnel stages, and has each one answered by an AI assistant that searches the web and cites its sources.
Each answer is read for topics, keywords, the brands and products it names, whether it recommends them, and its tone. Brand names are merged into one canonical name (and parent brand) and keywords are merged by meaning. Results are pooled over 14 days, with newer days counting more: today counts fully, a week ago half, two weeks ago a quarter.
- Mention share: the % of AI answers that name the brand.
- Citation share: the % of all sources AI cited that are about the brand or are its own site.
- Recommended: the % of answers naming the brand that recommend it.
- Trend: mention share this week compared with the week before.
- Step 9
Find the segments behind the questions
The pooled questions are clustered by meaning into up to ten user segments. The clustering settings are chosen by how cleanly the groups separate, and the page states how many questions fit a segment. Each segment gets a name, a description, traits, the needs it wants answered, and the topics and brands it asks about.
What does Category Overview show?
Four views per followed category: Questions, Keywords, Brands and products, and User segments.
Tracked questions
Rank, topic, intent, funnel stage, when it was first and last seen, trend and opportunity, with a mark when the question was verified on the web.
Question detail
The pain point, other ways people ask it, its keywords, the brands in existing answers, where it was asked and the source links.
Where the demand is
An intent × stage heatmap, topics by number of questions, and an opportunity map that plots demand against how poorly it's answered.
Keywords
Keywords by mentions, questions and topics, the top keywords, and a graph linking topics, keywords and brands.
Brands and products
Mention share, citation share, how often AI recommends each one, its week-on-week trend and how AI talks about it, plus the most-cited sites.
User segments
Up to ten segments with their size, traits and needs, a segment × topic view and who asks about what.
Who uses Category Overview?
Teams who need to know what a market asks before they decide what to say.
Brand and category managers
See which brands AI names and recommends in your category, how that moves week to week, and which sites shape the answers.
Content and SEO teams
Start from real questions with a demand proxy and an opportunity score, instead of a keyword list, and write for the gaps.
Product and insight teams
Read the pain points and segments behind the questions in Indonesia, Malaysia and Vietnam, in the language people used.
Category Overview vs keyword research vs social listening
They answer different questions. Keyword tools estimate how often a term is searched; social listening counts what's said about names you choose; Category Overview starts from the category and shows what people ask and what AI answers.
| Category Overview | Keyword research tools | Social listening tools | |
|---|---|---|---|
| Starting point | A category in a market | A seed keyword | Brands and terms you choose to track |
| Main unit | A real question, grouped with its paraphrases | A keyword | A post or mention |
| Demand signal | A proxy from how widely and consistently a question is found | Search-volume estimates | Mention volume |
| AI answers | Brands AI names, recommends and cites, and the sites it cites | Not covered | Not the focus |
| Why people ask | Intent, funnel stage, pain point and user segment for each question | Intent labels on some tools | Sentiment of posts |
| Refresh | Daily, with first-seen and last-seen per question | Varies by tool | Varies by tool |
Intura makes the product in the first data column, so read this comparison with that in mind. The other columns describe a typical tool in each category, not any one product.
What Category Overview doesn't tell you
Read the numbers with these in mind:
- Demand is a proxy built from how widely a question was found. It isn't search volume.
- Keywords, brands and segments come from AI answers to questions Intura writes each day, so they describe what AI answers say and cite, not what consumers post. Only the Questions view is built from questions people actually asked.
- Research reaches communities through web search, so private groups and logged-in feeds are out of reach.
- Coverage is the nine listed categories in Indonesia, Malaysia and Vietnam. If your category isn't on the list, it isn't covered yet.
- On the free plan you see a preview: the first three rows of each list. Charts are on paid plans.
Category Overview: common questions
What is Category Overview?
Category Overview is Intura's daily research on a product category in a market. It shows the real questions people ask about the category, which brands and sites AI answers name and cite, the keywords in those answers, and the customer segments behind the questions. You follow a category and the research refreshes every day.
Which categories and markets does Category Overview cover?
Nine categories: Beauty & Personal Care, Food & Beverage, Travel & Hospitality, Fashion & Style, Tech & Gadgets, Parenting & Family, Digital Products, Fitness & Health, and Finance & Investment. Three markets: Indonesia, Malaysia and Vietnam. A workspace follows up to three category-and-market combinations at a time, for example Fashion & Style in Indonesia and Food & Beverage in Malaysia.
Where do the questions come from?
From daily grounded web research through three lenses: communities and comments (forums, Reddit, Quora, Facebook groups, TikTok, YouTube and Instagram comments, marketplace Q&A), search (People also ask, autocomplete-style queries) and complaints and reviews. A question is kept only if at least two separate passes found it and it survives a web verification check.
How often is Category Overview updated?
Every day. Questions are tracked continuously with a first-seen and last-seen date and stay in the active view for 30 days after they were last seen. Brand, keyword and segment data are pooled over the last 14 days, with newer days counting more.
What do mention share and citation share mean?
Mention share is the percentage of AI answers that name a brand. Citation share is the percentage of all sources AI cited that are about the brand or are its own site. Both come from AI answers to about 100 questions per category and market each day.
How is Category Overview different from keyword research?
Keyword research starts from a term and estimates its search volume. Category Overview starts from the category and returns real questions, grouped with their paraphrases, with intent, funnel stage, pain point, a demand proxy and an opportunity score, plus what AI answers say about them. Its demand index is a proxy, not search volume.
How is Category Overview different from social listening?
Social listening tracks mentions of brands and terms you choose across social platforms. Category Overview doesn't need a list of names: it finds the questions a market asks, verifies them on the web, and shows which brands AI answers name, recommend and cite.
Is Category Overview on the free plan?
The free plan shows a preview with the first three rows of each list. The full research and every chart are on paid plans.
See what your market asks AI
Follow a category in Indonesia, Malaysia or Vietnam, and the first research runs the same day.
Free plan available · No credit card needed