Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.getcited.in/llms.txt

Use this file to discover all available pages before exploring further.

Not all queries are the same. A customer asking “what is GEO” has very different intent from one asking “best GEO tools for D2C brands.” Cited classifies every tracked prompt by intent type because the same brand can have very different visibility across intent categories — and the optimization work for each is different.

​
The four primary intent types

Cited uses a four-type taxonomy adapted from traditional search intent classification, with adjustments for how AI conversations differ from keyword searches.
Intent typeExample promptWhat the customer wantsWhat wins
Informational”what is mention rate”Education on a conceptReference content, definitional clarity, schema markup
Commercial”best HR software for Indian startups”Comparison and decision supportBrand presence in editorial coverage, third-party reviews, comparison content
Navigational”Keka HR login”Reach a specific destinationBrand authority, exact-match content, llms.txt
Transactional”buy noise cancelling earphones under 5000”Make a purchase decisionProduct pages, pricing transparency, purchase signals
Intent classification matters because the same brand can appear in informational queries (high mention rate, low conversion value) while being absent from commercial queries (low mention rate, high conversion value). Without intent segmentation, you cannot prioritize where to invest. A brand with 40% overall mention rate that scores 60% on informational prompts and 5% on commercial prompts has a different problem than one scoring 20% evenly across both.

​
How AI conversations differ from search queries

Traditional search queries are typically 2-4 words — “HR software India.” AI conversations are full questions, often multi-turn — “What HR software do Indian startups use? Which one is best for a 50-person company? How much does Keka cost?” This shifts the intent landscape in several ways:
  • More informational queries. People use AI to learn, not just to search. The conversational format invites longer, more exploratory questions.
  • More multi-step commercial queries. A single conversation can move from “best HR software” to “compare top 3” to “which is best for 50-person startup” — a funnel compressed into minutes.
  • Fewer navigational queries. Customers rarely ask an AI for a login URL. Navigation happens through bookmarks and search engines, not AI conversations.
  • Different transactional patterns. AI is used for research, with purchase happening on the destination site. The AI surfaces options; the customer converts elsewhere.
The Cited Index reflects this reality by using conversational, full-sentence prompts that mirror how customers actually talk to AI — not keyword-style search terms.

​
Why brands often misallocate effort

Most brands invest disproportionately in commercial intent — “best X for Y” queries — because that is where SEO budgets have historically flowed. But in AI search, informational intent is where reputation is built. The customer asking “what is content management software” today is the customer asking “best CMS for my marketing site” three weeks later. Brands that are cited in the informational answer become the brands customers consider when they shift to commercial intent. This upstream effect is measurable: brands mentioned in informational queries have higher share of voice on subsequent commercial queries in the same category. Ignoring informational intent means losing the upstream battle that determines downstream commercial visibility. The fix is not to abandon commercial queries — it is to invest in both, with awareness that informational content (glossaries, explainers, methodology pages) builds the foundation that commercial content capitalizes on.

​
Cited’s intent classification in practice

Cited operationalizes intent classification throughout the platform:
  • Every prompt in a brand’s prompt library is classified during query generation
  • Mention rate, share of voice, and average position are reported per intent type in the dashboard
  • Gap analysis surfaces prompts where competitors lead in a specific intent category
  • Recommendations are intent-specific — improving informational presence requires different content work than improving commercial presence
The Prompt Intelligence tab in the dashboard groups all tracked prompts by intent type, making it straightforward to identify where a brand is strong and where it has gaps.

​
Intent shifts within a single conversation

A single AI conversation can span multiple intent types. Consider this sequence:
  1. “What is content management software” — informational
  2. “What are the most popular ones” — commercial (comparison)
  3. “WordPress vs Webflow for a marketing site” — commercial (specific comparison)
  4. “How much does Webflow cost” — transactional research
Each step has different brand visibility patterns. Brands that win on step 1 (informational) often have an advantage on step 2 because the AI has already introduced their name in the conversation. This is the upstream-to-downstream effect that makes informational intent so valuable — it primes the customer before the high-intent question is even asked.

​
Frequently asked questions

Simplicity is a feature. More granular taxonomies (8 or 12 types) produce noisier classification and harder-to-act-on insights. Four types map cleanly to where brands need to invest differently — informational content, commercial comparison content, navigational discoverability, and transactional pages. Most brands cannot productively act on more than four categories at once.
During query generation, Cited’s classifier evaluates each prompt against pattern signals — question words, comparison language, brand specificity, and purchase intent markers. The classification is human-reviewable in the dashboard. Edge cases default to the most conservative interpretation.
No. Winning everywhere is rarely realistic. Most brands should pick 1-2 intent types where their position relative to competitors is weakest, focus there, and accept lower priority on the others. For most B2B SaaS brands this means commercial first; for most D2C brands this means informational first to build category presence.
Yes — the same intent classification is applied to prompts regardless of which platform they run against. What varies is platform behavior on each intent type. Perplexity tends to handle informational queries with citation-heavy responses; ChatGPT handles commercial queries with broader source mixing. The classification is platform-agnostic; the response analysis is platform-aware.