Topics and prompts

Overview#

A topic is an area of demand in your category, and a prompt is one realistic buyer question sitting inside that topic.

Between them they define everything Seerly measures. Visibility, sentiment, citations and competitor comparisons are all reported against your topics and all derived from the answers to your prompts, which means anything outside that set is not measured at all. It is not measured badly or partially; it simply does not exist as far as the product is concerned.

That is what makes this the highest-leverage thing in your account, and the part most worth getting right before you start reading numbers.

Interpreting results#

Topics come out of your category's search demand. Seerly maps the demand around what you sell, groups it into clusters and rates each cluster for how well it fits your brand, and during setup your topics are derived from that map together with the focus areas you entered. Afterwards the set is yours to manage: in the Keyword Universe you can promote clusters into tracked topics, which is how you add an area that Seerly did not pick up on its own.

Prompts are phrased the way people actually talk to AI assistants rather than the way they type into a search box, so "Best CRM for a small agency" rather than "crm small agency". They tend to fall into three recognisable kinds:

• Recommendation, along the lines of "what's the best X for Y?"
• Comparison, such as "X vs Y, which should I choose?"
• Experience, meaning "is X any good?" or "what do people say about X?"

Each kind exposes something different about your position, and it is entirely possible to be strongly present in one and completely absent from another. That is a specific and fixable problem, but only if you are looking for it, because a topic rank averages the three together and reports something unremarkable.

One topic, three kinds of prompt

Topic: project management for agencies

Kind	Prompt	You appear
Recommendation	What's the best project management tool for a small agency?	Yes
Comparison	Asana or Monday, which is better for a design agency?	No
Experience	Is it worth paying for project management software with five people?	Yes

The same prompts run every cycle, and that repetition is what makes the results comparable at all. When something changes you know the answers changed, rather than wondering whether you asked a slightly different question this time.

Best practices#

• Topics a colleague would recognise as a description of your business.
• Enough prompts per topic to be stable. With too few, results swing between runs on ordinary variation and you end up responding to nothing.
• A spread of prompt kinds rather than only recommendation questions, since those are the ones most brands are already reasonably good at.
• Country scoped to where you actually sell. A brand operating across several markets but tracked against one will only ever see one market's picture, and nothing will point that out.

Next steps#

• Review topics before trusting any metric. If they are wrong, fixing them is the single fastest way to make the entire dashboard mean something.
• Add prompts for questions you know buyers ask and that are missing from the set. You know things about your market that no extraction process will infer from your website.
• Remove topics you have no intention of competing on. They dilute your averages and consume prompt budget that would be better spent elsewhere.
• Watch the topic mix after a rebuild, since a refreshed map can shift what is tracked without anybody deciding that it should.

Limitations#

Seerly measures the prompts you track and nothing else. A question that no prompt represents is not being measured, which means gaps in your prompt set are indistinguishable from strength: nothing is reported either way, and silence reads as fine.

Prompt volume is capped by plan. See seerly.app/pricing.