Skip to content
Omnipresence

Prompts

The questions we ask on your behalf, and how often.

Updated October 20, 2026

A tracked prompt is a question a real buyer would type, asked on a schedule, on every assistant. Everything else in Omnipresence is downstream of this list, so it is the part worth spending an hour on.

A prompt is not a keyword

"crm software" is a keyword. "What CRM should a 12-person agency use?" is a prompt. Assistants answer the second kind, and the answer names a handful of brands. If your prompt list reads like a keyword export, you are measuring something nobody asks.

The three dials

  • Cadence. Daily for the prompts you are actively working, weekly for the rest. Cadence is what your model spend scales with.
  • Samples per run. Models answer the same question differently each time, so a measurement fires a batch of probes. More probes narrow the confidence band. Multiple samples per run are an Agency and Enterprise feature; below that a weekly cadence accumulates the same evidence over time.
  • Coverage. Every prompt runs on ChatGPT, Claude, Gemini and Perplexity, on every plan, asked the same way so the four can be compared.

Keep one prompt untouched

A control prompt never receives optimization work. When your tracked prompts climb and the control stays flat, the difference is attributable to the work rather than to the model drifting.

Concepts rather than clicks: why the same question gets different answers.

In this section

  1. 01Tracked promptsWhat a tracked prompt is, how to phrase one so the answer contains brand names, and how to keep a growing prompt set readable.
  2. 02Run cadenceChoose how often a tracked prompt runs, what each cadence costs on your own model key, and why changing it mid-measurement distorts a trend.
  3. 03Samples and confidenceWhy a measurement fires a batch of probes, what samples per run buys you, and how to read the confidence band on a Visibility Score.
  4. 04Control promptsA control prompt never receives optimization work. Why that makes it the only way to attribute a move to your work rather than to model drift.
  5. 05Topics and tagsGroup prompts into topics for per-topic visibility reporting, and use tags for the cuts that do not belong in the topic structure.
  6. 06Research promptsRun a set of prompts once for a study rather than tracking them: how research batches are budgeted, how to follow one, and how to pull the results.

Not tracking yet?

Seven days free, on your own model key. Point it at one brand and read what comes back.

Start your free trial