Prompts

A Prompt is a question that SurfacedBy sends to AI platforms during a scan. Each Prompt runs against every AI platform in your plan and the responses are scored for whether they cited your domain. Prompts are the unit of measurement behind every metric in the product.

How Prompts are generated

During onboarding, SurfacedBy analyses your domain to extract what you do, who you serve, and the category you compete in. It then assembles an initial Prompt set that mixes category-level questions (the kind a real buyer would ask), competitor-comparison questions, and Prompts derived from high-value search terms in your space.

The discovery logic is intentionally not exposed as a knob. That is by design: the value of the product is that it runs the Prompts that actually matter in your category, so we optimise that selection continuously rather than pushing the decision onto the user.

Curated versus user-supplied

Most users never touch their Prompt set. SurfacedBy pins a stable set over time and rotates out the ones that stop being useful.

If you want to be explicit, Tracked Prompts is the surface where you manage them. It has four tabs:

• Active - the Prompts scheduled on the next Full Benchmark, with status chips for winning, needs work, and at risk.
• Suggested by AI - new Prompts that SurfacedBy found in your category and thinks are worth adding. Pinning a suggestion moves it into Active.
• Matrix - a grid of where you rank on each AI platform, Prompt by Prompt.
• Inactive - paused or archived Prompts you can bring back.

On the Business tier, you can pin Prompts through the API using the POST /domains/{domain_id}/prompts endpoint. Pinned Prompts stay in the set across scans unless you unpin them.

Prompt counts by tier

Your plan determines how many Prompts are tested per scan. Counts vary by plan; see the Plans page for the current limit on each tier.