Setup
6 min readProjects, Competitors and Buyer-Intent Prompts
Everything Cited measures flows from three inputs: the project (your brand and domain), the competitor list, and the buyer-intent prompt set. Get these right and every audit answers a business question. This page covers how each works and how to maintain them as your market moves.
Projects
A project tracks one brand on one domain in one category. The project holds your competitor list, your prompt set, your audit history and your trend lines. Plans differ in how many projects you can run: Starter includes 1, Pro includes 2, Agency Pro starts at 3 client projects and grows with Scale Packs.
Project settings let you refine what Cited understood from your site: your offering description, your category wording, and your niche. Each of these shapes the prompts Cited generates, so a correction here improves every audit after it.
The niche field
If your brand serves a specific segment, say clean skincare for teens rather than skincare broadly, set it in the niche field in project settings. Prompts then ask what your segment's buyers ask, instead of generic category questions your buyers never type. This matters most for D2C and consumer brands in broad categories.
Your competitor list vs the category benchmark
These are two different things, by design:
- Competitor list: The brands you choose to track closely. They get named columns in comparisons, threat alerts, and head-to-head trend lines.
- Category benchmark: Every brand the AI engines actually named on your prompts, whether you listed it or not. AI answers routinely include brands outside anyone's named competitor set, and unlisted brands winning your prompts is a signal you want surfaced, not filtered out.
If an unlisted brand keeps appearing in your answers, add it to the competitor list from the results screen to start tracking it closely.
Second-order effect worth knowing
The brands your sales team calls competitors and the brands AI engines put in your answers overlap less than most teams expect. The gap between those two lists is usually the fastest route to a strategy conversation leadership has not had yet.
Buyer-intent prompts
Prompts are the questions Cited asks the engines on every audit. Cited generates them from what it learned about your offering, spread across personas and funnel stages. Prompt depth per audit is set by plan: 8 on Free, 20 on Starter, 40 on Pro and Agency Pro.
The set deliberately mixes intent types, because each maps to a different revenue moment:
- Discovery: A buyer describes the problem, no brand names: best way to handle X, tools for Y. These measure whether engines think of you at all.
- Comparison: A buyer weighs options: X vs Y, alternatives to Z. These measure how engines frame you against rivals.
- Switching: A buyer wants out of a current tool or product. These often decide deals and are the most overlooked.
How stage-level results turn into stage-level fixes is its own discipline: see Closing the Loop on Buyer Intent and Funnel Stages.
Editing and maintaining the set
Edit prompts during onboarding or any time in project settings. Changes apply from the next audit. Two rules keep the data useful:
- Write the way buyers talk: Full questions in natural language, not keywords. Buyers ask engines in sentences, and the answers differ.
- Keep discovery prompts unbranded: A discovery prompt containing your brand name measures nothing: the engine will mention you because you asked it to. Save brand names for comparison and switching prompts.
Keep the set stable while you are measuring the effect of fixes: changing prompts and content in the same window makes it hard to attribute movement. Revisit the set quarterly, or when you launch into a new segment.
Related docs