Prompts Tab: User Guide
Prompts are the questions AnswerRank asks AI models on your behalf. This page shows your results for each question separately, so you can see which ones you show up for and which ones you don't.
What you see on the page
- Filter bar: date range, Models, Countries, Tags, plus Export (top right) to download the table.
- Prompts / Suggested Prompts switch. *Prompts* is the list being tracked. *Suggested Prompts* stays empty until you click Generate More Prompts; AnswerRank then proposes new questions worth tracking.
- Add Prompt (green button): add your own question to track.
- Prompts table: one row per question, with sort, search ("Filter prompts") and a Columns picker. The footer shows two counts: the number of questions, and the total of questions multiplied by AI models.
- Arrow next to a row: expands it into one sub-row per AI model (Perplexity, Google AI Overview, Gemini, ChatGPT), so you can see which engine is helping or hurting you.
- Summary (bottom): AnswerRank's own description of your brand, what it offers and who it competes with. If it is wrong, your prompts and competitors may be off too.
How to read the columns
| Column | What it means |
|---|---|
| Visibility | % of answers to this question that mention you |
| Sentiment | How positively you are described in those answers |
| Rank | Your average position among brands named (lower is better) |
| Market Share | Your share of all brand mentions in answers to this question |
| Model | Logos of the AI engines this row covers |
| Country / Tags | Market the question is asked from, and your own labels for grouping |
Prompt Details page
Click any per-model sub-row to open the drill-down for one question on one AI model: - Header: the question, created date, model, country and tags. - Performance: a Ranking bar chart of which brands that model mentions most for this question, and a Trends line chart of each brand's daily visibility since tracking began. Switch between Visibility, Rankings and Sentiment. - Insights → Statistics: Total Responses (how many times the question has been answered), total mentions of your brand, and the Top Competitor with its mention rate. - Insights → Response History: the actual AI answers, one per day, with coloured chips for every brand named. Read a few to understand *why* you are or aren't mentioned. - Insights → Sources: websites the model drew on for this question, with category and number of mentions. - Insights → Direct Links: specific pages the model linked to, with counts.
What matters most
- Prompts with low Visibility. Questions where AI rarely or never mentions you are your clearest gaps.
- Prompts with high Visibility but low Sentiment. You are mentioned, but not in a good way.
- The per-model breakdown. A 0% on one engine next to good scores on the others usually means that engine relies on different sources.
Tips
- A good way to start: sort by Visibility from lowest to highest, open the weakest questions and check which model is pulling the number down. Then read that model's answers in Response History.
- Add tags to group questions by topic or product, then filter by tag to see how each topic is doing.
- Read the Summary at the bottom of the page. If it describes your brand wrongly, fix the project setup first, because everything else depends on it.
- Use Generate More Prompts to find questions you hadn't thought of, but only add the ones real customers would ask.
- Try not to add or remove lots of prompts at once. It changes all the averages and makes trends hard to read.
- In Prompt Details, the Sources and Direct Links tabs show which websites and pages you would need to appear on to win that question.
Export
Yes. Click Export (top right) and choose Download CSV or Download Excel. You get one row per prompt with its metrics for the selected period. The file uses the filters that are currently applied (period, models, countries, tags), and Edit next to Columns lets you pick which of the 16 columns to include.