Monitor
Prompts
Add, generate, track and pause the questions Ranqo asks AI engines for your brand, and read each prompt's results on its own page.
A prompt is a question Ranqo asks each AI engine on your behalf, in the words a buyer would use. Every visibility number starts here: a tracked prompt is asked on every weekly run, on every engine it is set to, and each answer is read for your brand, your competitors and the sources cited. Open the page from Prompts in the Monitor section of the sidebar.
The three tabs
- Library: the set of prompts you ask, with the controls to add, track, pause and delete them.
- Themes: topic clusters from research on your category, each one a source of new prompts. See Themes.
- Tracking: what came back for each tracked prompt. See The Tracking tab below.
The tab you pick is kept in the page address, so a link or a reload opens the same tab.
Tracked, Paused and Never run
Every prompt is in one of three states, shown by the switch in its Tracking column:
- Tracked: asked on every weekly run.
- Paused: was tracked before, but sits out the weekly run. It keeps whatever history it has.
- Never run: has never been tracked, so it has no results yet.
Flip the switch to track or pause a prompt. Every change ends in a message with Undo. Undoing the tracking of a prompt that had never run returns it to Never run.
The card's subtitle says what the next run will do. It gives the next run's date, or says that weekly runs start once at least one prompt is tracked, or that tracking is paused for the whole brand. A brand is paused from the Pause button on its Home page.
Pausing your last tracked prompt
Weekly runs need at least one tracked prompt, so pausing the last one asks first: Pause your last tracked prompt? Everything measured so far stays, but no run happens until you track a prompt again. Choose Pause anyway to go ahead.
When nothing in the library is tracked, a banner offers Track all N (or Track this prompt when there is only one) to switch every prompt on at once.
Adding prompts
The two buttons in the card header add prompts. Both are disabled once you reach your plan's prompt limit, with a note saying how many of your plan's prompts you have used.
Add manually
Add manually offers Write one or Import a list:
- Write one: type the prompt, pick its category, the engines to ask it on (Track on) and the market it targets.
- Import a list: paste one prompt per line. The category, engines and market you pick apply to all of them. Lines already in your library are skipped.
Track on and the market start with the brand's settings. Each prompt keeps the engines and market it is saved with, and every run asks it on those engines, from that market. Changing the brand's market later leaves existing prompts as they are, and changing the brand's engines updates only the tracked prompts whose engines matched the brand's. See Brand settings. To change one prompt, choose Edit from its ⋯ menu.
Prompts you add by hand are tracked from the next run.
Generate with AI
Generate with AI opens the Generate prompts page:
- Your first set. With no prompts yet, Ranqo writes 10 starter prompts from your saved brand profile, spread across the unbranded categories. Choose the market and the engines, then press Generate prompts.
- Later sets. You can add an optional Description of what to focus on, and choose the market, the engines (Track on), the Categories and the Number of prompts. The number stops at the slots left on your plan. Prompts that repeat ones you already have are skipped.
Generated prompts are saved to your library as Never run. Tick the ones you want and press Add to Tracking. When those are the brand's first prompts, adding them starts your first analysis right away and takes you to the Visibility page.
Branded and unbranded prompts
An unbranded prompt asks about your category without naming you, the way a buyer who has not heard of you would ask. A branded prompt names your brand.
Ranqo writes unbranded prompts by default. A prompt that names your brand gets an answer about your brand almost by definition, so a visibility score built on branded prompts says little about whether new buyers find you. The two categories that can name you, Comparison and Brand Research, are left out of generation unless you select them, and selecting them shows a warning.
A prompt marked as branded carries a Branded pill in the library. The All Prompts, Branded and Unbranded filter on the Visibility and Competitors pages separates the two.
The Library table
| Column | What it shows |
|---|---|
| Prompt | The question. Click it to open the prompt's insights page (not available for a prompt that has never run). |
| Category | Discovery, Problem/Solution, Use Case, Expert/Deep Dive, Comparison or Brand Research. |
| Volume | The writing model's estimate, made once when it wrote the prompt, of how often your audience asks this kind of question. Not a search count. Empty for prompts you add yourself. |
| Difficulty | How crowded the answers are with competitors, from no competition to extreme. |
| Platforms | The engines the prompt is asked on. |
| Location | The market the prompt targets. |
| Added | When the prompt joined the library. |
| Tracking | The switch and the state word. |
Difficulty starts as an estimate made when the prompt is written. After every run that asks the prompt, it is replaced by a measured level based on how many competitors the answers named. An empty meter means either that the answers named no competitors or that there is no estimate yet; hover it to see which. When the model that wrote a prompt gave no estimate, the middle level is stored instead, so the prompt shows High competition until its first run. See Difficulty and volume.
Sort by Volume, Difficulty or Added. Filter with the search box and All Categories, All Themes, All Platforms, All Locations, All Tracking (the three states) and All Dates (the month a prompt was added). The bar above the filters shows how the visible prompts split across categories.
Acting on several prompts
Hover a row's number to reveal its checkbox. Once rows are selected, a bar offers Track selected, Pause selected and Delete…, and a checkbox appears in the header that selects every prompt the current filters show.
Editing and deleting
The ⋯ menu on each row offers View insights, Edit and Delete….
Deleting always asks first, because a deleted prompt's answers go with it and cannot be recovered. Figures worked out from those answers, past runs included, can change. When the prompt is tracked, the dialog offers Pause instead, which takes it out of the weekly run and keeps its history.
The plan meter
The footer counts your prompts by state and shows how many of your plan's prompts you use, in the form N of M prompts on your plan. The allowance is shared across all your brands. It counts every prompt, tracked, paused or never run, so pausing a prompt does not free a slot. To free one, delete a prompt. An Upgrade link appears as you near the limit. See Plans and limits.
The Tracking tab
The Tracking tab shows what came back for each prompt in the last 30 days, or the last 7 if you pick that range (a longer range still shows 30 days):
- Visibility: the share of the prompt's answers in the window that named your brand.
- Position: where you appear when named, averaged over the window. Lower is better.
- Mentions: answers that named you, out of every answer in the window.
- Engines: one logo per engine. A logo keeps its colour when that engine's latest answer named you and is dimmed when it did not. Hover a logo for its rank and its count across the window.
- Top Competitors: the brands AI named most often for this prompt.
The bar above the table groups the prompts by how often they win a mention, in the same score bands as the visibility rings, plus Never named for prompts asked but never answered with your brand. The All Platforms filter recalculates every column for the engines you pick. When the window holds no run, an As of note gives the date of the run the figures come from. A query that failed counts here, and on a prompt's insights page, as an answer that did not name you; see When a query fails. Click a row to open the prompt's insights page.
A prompt's insights page
Each measured prompt has its own page, reached from its row or View insights:
- The header shows the prompt with its category, market, theme and tracking state.
- All Platforms and the date range narrow the page.
- Visibility, Position and Sentiment, each with its change against the previous period, and Sources: the distinct sites cited, with the share of citations whose source mentions your brand.
- Volume and Difficulty meters.
- The trend chart, one point per run, with Sentiment on the same 0 to 100 score as the card, then Performance by Platform, Brand Rankings for this prompt, Citation Sources and AI Response History, where you can read every answer the engines gave.
A prompt with no answers yet shows Nothing measured yet. When it is tracked, the page gives the date of the next run. When it has never run, the page offers Track this prompt.