How do you track our personas?

One persona per prompt

The unit of tracking is a prompt, and each prompt belongs to exactly one persona. That is what makes the split clean. A question about total cost and contract terms is, for example, an economic buyer's question; one about integrations and workflow is a champion's; one about permissions and provisioning is an admin's. When the daily runs put those prompts to the engines, every answer is already labeled with the role that would have asked it, and nothing has to be inferred afterwards.

The CFO's shortlist and the admin's shortlist come from different questions, so they can come out different. You can win one and never appear in the other, and this is where that shows up.

What the persona views show

The mention rate per persona shows where you are shortlisted and where you are absent, with the change since the previous period. Filter share of voice to a single persona and you see who takes your place on that role's shortlist, which is the name the loss was missing. The same filter applies to every other view, so the prompts, the full answers, and the cited sources can all be read for one role at a time. Your strategist works from the same views your login opens.

Defining the personas

We define the personas with you at the start of the engagement, from the buying committee you already sell to, rather than leaving a model to guess them from the prompts. The tracked question set is then built to cover each role, so no persona is represented by two prompts while another has twenty.