Skip to main content
Use Groups for collections of related questions and Personas for hypothetical audiences. Both organise your monitoring programme; a persona assignment does not rewrite the question sent to an AI platform.
Group comparison chart for branded and unbranded prompts

Compare prompt groups over the same period. Example workspace.

Create groups around a product category, market, customer need, or campaign. Assign prompts in the prompt editor or through the available bulk actions. Use descriptive names such as “UK accounting comparisons” rather than a label that only one team member understands. Keep branded and unbranded research distinguishable. A group of questions naming your brand should not be interpreted as a neutral test of category discovery.

​
Define an audience persona

Create a persona with a name and occupation, then add relevant profile details. Assign the prompts that belong to that audience. A persona is an organisational description, not a real customer profile or evidence of how that audience behaves. If audience context should influence the answer, put that context in the question itself. For example, “Which accounting tools suit a freelance designer?” explicitly asks about that audience. Attaching a persona alone does not add those words to the monitored prompt.

​
Use the same structure in integrations

The public API calls prompt groups tags and exposes them with tag_ identifiers. It exposes persona assignments with per_ identifiers. These represent the same workspace organisation used in the dashboard, so integrations can preserve your grouping when creating or exporting prompts.