Skip to main content
Prompt Research is where you discover new AI-query opportunities before you start tracking them. Instead of manually brainstorming prompts, you organize research into themes — topic clusters that group related queries together — and let Mentionpath generate the individual queries within each theme. Each query is scored for search volume, competitive difficulty, AI likelihood, and opportunity, so you can quickly identify which ones are worth tracking and which to deprioritize. The workflow moves from broad to specific: create or accept a theme, generate queries within it, review the scores and status labels, and add the most promising queries directly to your tracked prompt list with one click.

​
Themes

A theme is a topic cluster — a named grouping of related AI queries. For example, a theme called “Onboarding for SaaS” might contain queries like “how to onboard new users in a B2B SaaS product”, “best onboarding tools for SaaS”, and “what makes a good SaaS onboarding experience”. You can create themes manually or accept AI suggestions.

​
Creating a theme manually

Click Create Theme on the Prompt Research page. Enter a name for the theme and an optional seed — a one-line description of the core topic that guides query generation. The seed is not shown to end users; it’s used internally to steer the AI when it generates queries for the theme.

​
AI theme suggestions

Below the themes table, Mentionpath shows AI Theme Suggestions — theme ideas generated based on your domain context, tracked prompts, and competitors. Suggestions are grouped into three segments:
Topics where AI models rarely or never mention your brand, but competitors have established presence. These represent gaps in your AI visibility — areas where you need to create content and build authority to become part of the conversation.Accepting a blind spot theme is a strategic move: you’re identifying an area where competitors already appear in AI answers and you don’t. Closing that gap requires both content creation and prompt tracking.
Topics where you have some existing presence or natural advantages, but your AI visibility doesn’t yet reflect that strength. You’re on the map, but not as prominently as you could be.Opportunity themes are often the highest-priority targets: the groundwork is already there, and focused content or prompt optimization can translate existing authority into AI citations.
Topics where you already perform well in AI-generated answers. These themes are worth monitoring rather than urgently expanding — you want to maintain your position and watch for competitors gaining ground.Accepting a strong presence theme lets you set up ongoing tracking so you’re alerted if your visibility starts to slip.
Each suggestion card shows the theme name, a seed description, and a brief explanation of the opportunity. Click Research Theme to accept the suggestion and add it to your themes list. When you accept a suggestion, Mentionpath starts generating queries for that theme in the background. Click Regenerate at any time to get a fresh set of AI suggestions. Individual suggestions you’re not interested in can be dismissed without affecting others.

​
Exploring queries within a theme

Click any theme row to open the theme detail view. This is where you see all the individual queries Mentionpath has generated for that theme, along with scoring data for each one. If the theme has no queries yet, click Generate Queries to start AI query generation. On subsequent visits, click Generate More to add additional queries.

​
Query fields

Each query row contains the following information:
The search intent behind the query — what the person asking it is trying to accomplish.
IntentDescription
LearnInformational queries. The person wants to understand a concept, process, or topic.
ConsiderEvaluative queries. The person is comparing options and building a shortlist.
PurchaseCommercial queries. The person is close to a buying decision and looking for final validation.
Use the All Intents dropdown to filter the query list by a single intent.
A more specific classification of the query format, which reflects how you’d need to structure content to answer it well.
TypeExample format
How-toStep-by-step instructional content
SetupConfiguration or installation guidance
ExplainerDefinition or concept explanation
Best”Best [X] for [use case]” roundup
CompareHead-to-head feature comparison
Alternatives”Alternatives to [product]” listicle
PricingPricing breakdowns and cost comparisons
ReviewsUser or expert review formats
Case StudyReal-world outcome narrative
Selection CriteriaWhat to look for in a solution
Due DiligenceVendor evaluation framework
DemoProduct demonstration or walkthrough
MeasurementMetrics, KPIs, and how to track them
ImplementationDeployment and rollout guidance
Click the Type column header to filter the query list by a specific query type.
Both volume and difficulty are shown as five-segment bar indicators rather than raw numbers, making it easy to compare queries at a glance.Volume (1–5): An estimate of how many people search for this type of query. Higher volume means more potential reach. Green bars indicate high volume, orange indicates moderate, and red indicates low.Difficulty (1–5): The competitive difficulty of ranking for this query in AI-generated answers. Lower difficulty means an easier path to visibility. Green bars indicate low difficulty (good), amber indicates moderate, and red indicates high difficulty (harder to break through).You can sort the query table by volume or difficulty by clicking the respective column headers.
How likely it is that AI models (such as ChatGPT, Gemini, or Perplexity) will generate an answer for this type of query. Queries with high AI likelihood are the most valuable targets for optimization — they’re the queries where showing up in an AI answer gives you the most reach.
A composite score that weighs volume, difficulty, and AI likelihood together to produce an overall opportunity rating. Queries with high opportunity scores represent the best combination of reach, achievability, and AI relevance.

​
Status labels

Once scoring is complete, each query receives a status label that summarizes its overall strategic position.
High volume, manageable difficulty, and strong AI likelihood. These are your best opportunities — prioritize creating content and tracking these queries.
Decent signals across volume, difficulty, and AI likelihood, but not a standout in any single dimension. Worth tracking to gather data and see how your visibility evolves.
High volume but also high difficulty. Reaching visibility here is possible but requires strong, differentiated content. Consider your competitive positioning before investing.
Your brand already appears in AI answers for this query. These are wins to maintain and monitor rather than actively pursue — but tracking them confirms your presence is holding.
Low volume, low AI likelihood, or both. Not a priority for most brands. May be relevant for very niche audiences or long-tail strategies.
Scoring happens automatically after query generation. While scoring is in progress, the status column shows a “Researching” indicator. You don’t need to stay on the page — scoring runs in the background and results appear when you return.

​
Adding queries to your tracked prompts

When you find a query worth monitoring, click Track in the actions column. Mentionpath adds the query to your tracked prompts list and begins monitoring how AI models answer it and whether your brand appears. Once a query is tracked, the Track button is replaced with a “Tracked” confirmation. You can filter the query list by status to find queries that are already tracked, still being scored, or awaiting your decision.

​
Filtering and searching

Within a theme’s query list, you can:
  • Search by keyword — filter to queries containing specific text
  • Filter by intent — show only Learn, Consider, or Purchase queries
  • Filter by query type — show only a specific format (e.g., Comparison or Pricing)
  • Filter by status — show only Promising, Competitive, Already Visible, or other labels
  • Sort by volume or difficulty — click the column headers to cycle through ascending, descending, and unsorted
All filters can be applied together and cleared independently.
Start with Blind Spot themes suggested by the AI — these are the areas where your competitors are already getting cited and you’re not. Generating and tracking queries in these themes gives you the fastest path to closing visibility gaps.