Ask questions across your CRM, trade jobs and knowledge base.
Follow these steps
1
Open the assistant
Go to Content in the sidebar — the assistant opens with no conversation selected. Send your first message and a conversation is created for you automatically.
2
Ask across your workspace
Ask about a lead, a trade job, a product, or anything in the knowledge base, in plain language. If you’re referring to something already in this conversation — "make it shorter", "the LinkedIn one" — say so; the assistant looks up what you mean before it answers.
3
Watch it work, not just wait
Replies stream in token by token, and when the assistant needs to look something up or take an action, you’ll see that tool call appear in the conversation as it runs, followed by its result.
4
Pin what it should remember
Attach a product, an audience, or other context to the conversation so the assistant keeps it in view for every later message, instead of you repeating yourself each turn.
5
Come back anytime
Every conversation has its own address, so refreshing the page, using the browser’s back button, or opening it again from History all return you to exactly where you left off.
Each message uses an AI credit
Sending a message consumes one of your organization’s AI credits before the assistant starts replying. If you’re out of credits, the assistant tells you instead of sending a partial answer.
Screenshot pending — Assistant conversation streaming a reply with a tool call in progress
A response streaming in, with a tool call shown as it runs.
Screenshot pending — Conversation history list alongside pinned items for the active conversation
Past conversations and pinned context for the one you have open.
Want the full detail?
This guide covers the day-to-day flow. For the streaming architecture, tool access, and exactly what the assistant can see, see the AI Chat / Assistant doc.