Skip to content

Ask Vercatus Intelligence

Ask in the top bar (or ⌘ J, Ctrl J on Windows) opens Vercatus Intelligence, the assistant. Ask it anything about your workspace or about Vercatus, in your own words and your own language: “What should we do first on our main site?”, “Why did Perplexity stop naming us?”, “What does Answer Share count?”, “How do I connect Search Console?”. It answers in plain language and links to the screen or the documentation page behind the answer.

  • Your workspace, as you see it. The assistant reads your sites and what is measured on them, runs, competitors, plays, results, setup and plan through the same access you have. It cannot see anything your role cannot open, and it never sees another workspace.
  • The public documentation, the published knowledge base and the product’s own pages.
  • Where you are. The screen you asked from and the site chosen in the top bar go with the question, so “this site” means the one you are looking at.

It never sees passwords, API keys, tokens or webhook addresses: those are removed before anything is read.

The assistant explains and points; it does not act. It does not start runs, change settings, invite people or change plays. When something needs doing, it tells you how and links to the place to do it. Answers can be wrong: check the screen they link to before you act on a figure.

Every conversation is kept under Your conversations in the assistant, with its answers.

  • Closing. Close a conversation when you are done. One without a new message for 24 hours closes by itself. A closed conversation cannot continue; start a new one.
  • Rating. Once closed, say whether it solved your question and rate it from 1 to 5, with a comment if you like.
  • A person. Ask for a person in the conversation, or press Talk to a person. The Vercatus team gets the conversation and replies by email.

Your conversations are stored in your workspace. The Vercatus team reads conversations that were rated low, marked not solved, handed to a person, or that the assistant could not answer, to reply and to improve the answers. Questions and the data the assistant reads to answer them are processed by our model provider under no-training terms, like the rest of the analysis (see Security and data handling).

A person can ask the assistant 30 times an hour, and a workspace a number of times a day that comes with the plan: 30 on Starter, 100 on Growth and Portfolio, 200 on Enterprise. A conversation holds up to 60 messages; start a new one after that to keep answers sharp.