Discover

Run prompt discovery across your website, search data, social signals, and competitor gaps to find the prompts worth tracking

Find the prompts your market is using - and the gaps your brand isn’t covering yet - from one discovery surface.

Overview

Discover is Athena’s prompt discovery engine, available as the Discover tab on the Prompts page. It scans real data sources - your website, Google Search Console, social conversations, and your keyword landscape - then produces a structured, ranked set of suggested prompts you can review and start tracking. It replaces the earlier Planner and Universe tabs; old links redirect here automatically.

You need access to Discover on your plan. Individual sources can have separate access requirements; unavailable sources show Requires an upgraded plan. Search Console also requires a connected Google Search Console property.

What it does

Maps your category - Analyzes your website, segments, and products, plus the search terms and community discussions in your space.

Generates real prompts - Translates raw search data and audience conversations into prompts a real user would type into a search engine or AI assistant.

Checks existing coverage - Compares proposed questions with the prompts you already track and removes redundant suggestions. A question that asks buyers to discover a suitable product serves a different purpose from a question evaluating a named product, so both may be useful. Already tracked labels identify suggestions already in your monitored set.

Makes it actionable - Select the suggestions you want and click Start tracking to add them to your monitored prompt set.

Sources

When you start a run, you pick which sources to scan:

• Website analysis - market analysis from your website, segments, and products.
• Search Console - clusters your Google Search Console keywords into prompts grounded in real search data. Requires GSC to be connected in Settings > Integrations; otherwise the source appears disabled with a connect link.
• Social - Reddit threads, YouTube comments, and videos in your category.
• Keyword gaps - customer questions informed by competitor search keywords. These indicate market demand; they do not prove that an AI engine currently cites a competitor or excludes your brand.
• Custom instructions - describe the topics, products, audience, or intent you want the run to focus on. Selecting this source requires a written brief. Your instructions guide the run; they are not treated as verified product facts.

Athena also reads your website to connect the selected market signals to your products and services, even if Website analysis was not selected. The run’s source summary explains when website evidence was added for this purpose. Connecting Search Console is optional unless you want to use that source.

Running a discovery

1. Go to the Prompts page and click the Discover tab.
2. Click Run discovery.
3. Name the run, choose a country and language, and check the sources you want to scan.
4. Click Run discovery. Sources run in the background; completion time depends on the selected sources and available evidence.
5. Open the run from the history list to follow progress and review results. Check its status and source summary before interpreting the output; a failed or empty source means that evidence was unavailable for the run. You can leave the page while it runs.

Writing useful custom instructions

Name the products, customer needs, and market you want to cover. Include exclusions and say whether you want branded questions, non-branded discovery questions, or both. You can include relevant product URLs to clarify the scope.

For example: “Focus on organic bedding for people who sleep hot. Cover material choices, care, and durability. Include non-branded shopping questions and branded product comparisons. Exclude bath products.”

Examples communicate the direction of the run, so use them as illustrations rather than a fixed list to reproduce. Avoid asking Athena to assume product claims that your website does not support. Market signals establish what people need; your website establishes which needs your products can address.

Reviewing results

Suggestions are grouped by Topic in a spreadsheet-style table with columns for:

• Source - which source produced the prompt
• Funnel - TOFU (awareness), MOFU (consideration), or BOFU (decision)
• Intent - Informational, Commercial, Transactional, or Navigational
• Type - Head, Mid tail, or Long tail
• Volume - estimated monthly search volume
• Value - estimated keyword value (CPC)

Signal chips (such as Competitor Gap and Social Signal) appear above the table when the run found prompts carrying those signals - use them to surface the strongest opportunities first. Hover the info icon on any row to see why the AI suggested that prompt.

Select prompts individually or by topic, then click Start tracking (N). Athena checks for duplicates and shows a count of new vs. already-tracked selections before confirming. You can also export the currently filtered view as a CSV.

When to use it

You’re setting up tracking for a new brand or category - Run discovery to seed your monitored prompt set with a defensible, data-backed starting list rather than guessing.

You want to explore competitor demand - The Keyword gaps source uses competitor search keywords to find relevant customer questions. Use your tracked response and citation data separately to measure who actually appears in AI answers.

You want to expand into a new product line or audience - Run a new discovery with custom instructions scoped to the new area before you commit to a tracking plan.

You’re filling gaps in an existing prompt set - Re-run discovery periodically - every few months, or when you expand into a new product area - rather than just once at setup.

If a run returns few or no prompts

Review the run’s source summary. Could not collect evidence means a source failed; No usable evidence means it returned nothing usable. A small result set can reflect narrow instructions, limited supporting evidence, or coverage already in your tracked prompts. It does not by itself prove that you have covered the entire market.

Check the website content and source connections, broaden overly narrow instructions, and start a new run when needed. Earlier runs remain in history and may show the older discovery layout. See the Discover run reference for the current result controls.