Skip to main content
Prompt tracking is how VeluraHQ continuously samples AI engine responses to understand your brand’s visibility. You define prompts that mirror real user questions; VeluraHQ runs them across AI platforms on your chosen schedule and stores every response for analysis. Over time, each prompt builds a history of how AI engines have answered that question — which brands they mentioned, which pages they cited, and how those patterns have shifted as models update and competitive content changes.

​
What is a prompt?

A prompt is a structured question or query that simulates a real user search. Examples include “What are the best enterprise SEO workflow platforms?” or “How do I refresh content for AI Overviews?” or “Which tools help content teams measure AI search visibility?” VeluraHQ submits these prompts to each AI platform you have configured and records the full response — which brands are mentioned, which URLs are cited as sources, how high in the answer each brand appears, and how the brand is described. Every execution is stored as a run, giving you a time-stamped record you can compare across weeks and months. The more specific and realistic your prompts, the more useful the results. Write prompts the way your target audience would actually phrase a search query, not the way you would describe your own product.

​
Creating prompts

1

Go to AI Search Analytics > Prompts

Open the AI Search Analytics section from your workspace sidebar, then select the Prompts tab.
2

Click Add Prompt

Select Add Prompt to open the prompt builder.
3

Enter the prompt text

Write the prompt as a real user would phrase it. Avoid using your brand name in the prompt itself — neutral, category-level questions give you the most accurate read on organic visibility. For example: “What platforms do enterprise content teams use for SEO workflows?” rather than “Does [Your Brand] help with SEO workflows?”
4

Assign a topic category

Choose or create a topic category to group related prompts together. Categories keep your prompt library organized and let you filter results by theme — for example, “Content Refresh,” “AI Visibility,” or “Competitive Positioning.”
5

Select target AI platforms

Choose which AI platforms to run this prompt against. You can select all platforms available on your plan or target specific ones where you want focused visibility data.
6

Save the prompt

Save the prompt. It is now available to run manually from the Prompts tab or to include in a scheduled run. No further configuration is required before your first execution.

​
Scheduling prompt runs

Rather than running prompts manually each time, you can put them on a schedule so VeluraHQ collects data automatically. Scheduled runs catch competitive shifts before they cost you traffic — a competitor publishing a strong piece of content or an AI model update can change citation patterns overnight. To schedule a prompt, open it from the Prompts tab, select Schedule, and choose a frequency:
  • Daily — best for high-priority topics where you need near-real-time competitive intelligence
  • Weekly — recommended for most brand and category prompts; balances coverage with credit efficiency
  • Monthly — suitable for lower-priority topics or exploratory prompts where trends matter more than timeliness
Example schedule setup: A content team tracking five core topic areas might schedule their primary brand awareness prompts daily, their category comparison prompts weekly, and their exploratory long-tail prompts monthly — giving them continuous coverage of what matters most without burning through credits on lower-signal queries.
Run prompts more frequently for topics where you know competitors are actively publishing. If a rival just launched a new product or published a major piece of content, switching relevant prompts to a daily schedule for two to four weeks lets you measure the citation impact in near real time.

​
Reading prompt results

After a prompt runs, open it from the Prompts tab to see the full results for that execution. Each run report includes:
  • Brand mentions — which brands appeared in the AI’s response and how many times each was referenced
  • Cited URLs — the specific pages the AI engine linked to as sources, including which domain they belong to
  • Brand position — the order in which your brand appeared in the answer relative to competitors
  • Sentiment and description — how the AI characterized your brand, including the language and framing it used
  • Trend data — how the current run compares to previous executions of the same prompt, so you can see direction of travel at a glance
Click into any individual run to read the full AI response with brand mentions and citations highlighted inline.

​
Prompt history

Every execution of a prompt is stored in full — not just the summary metrics, but the complete AI response. This gives you a time-stamped archive you can return to whenever you need to understand how visibility has changed. Use the timeline view on any prompt to see a chronological record of all executions. The timeline makes it easy to identify exactly when a competitive shift happened — for example, when a competitor first started appearing in a response where they previously had no presence, or when your citation rate dropped after an AI model update. You can compare any two runs side by side to see what changed in the AI’s response between them.

​
Analyst assistant

The built-in analyst assistant lets you ask plain-language questions about your prompt data without needing to build reports manually. The assistant grounds every answer in your actual workspace data — it does not generate generic advice. Example questions you can ask the analyst:
  • “Which prompts lost the most citations this month?”
  • “Which AI platform gives us the lowest average position?”
  • “Show me the prompts where Competitor A is cited and we are not.”
  • “Which topic categories have improved the most over the last 30 days?”
The analyst returns direct answers with supporting data, and you can use its output as the starting point for a content action or workflow run.
Each prompt execution against an AI platform consumes credits from your plan. If you have a large prompt library, batch your scheduled runs during off-peak hours to maximize efficiency. You can review your credit consumption in Workspace Settings > Usage.