The activity feed: everything that ever happened, in order
Chats, calls, AI tasks, meetings, stage changes, assignee changes, comments and captured form submissions on one filterable feed.
The activity feed is the answer to "what did we tell them in June" and to "why is this customer annoyed", and both answers are usually one scroll away.
What makes it work is that everything lands on it. Not just conversations — calls with their outcomes, tasks the AI agent created, meetings, form submissions, and audit entries for stage and assignee changes. Nothing in the product edits a contact silently.
Steps
Open a contact, or read it beside a conversation
It is the same feed either way — the contact's history, not the conversation's.
Know what lands on it
Chats, calls, AI tasks, meetings, stage changes, assignee changes, comments and captured form submissions.
Read call entries for the outcome, not the fact
"Call · 4:32 · Reached — spoke" tells you something. Four "no answer" entries in three days tells you a great deal more, and should stop you dialling again today.
Read audit entries to answer "who changed this"
Stage changes and assignee changes are entries rather than silent edits. "Qualified → Won · by Priya Nair" is the whole answer to a question that is otherwise unanswerable.
Filter it when the history is long
An all-activity filter sits at the top. On a customer of two years, this is the difference between finding something and giving up.
Compose from the top
Comment, task and meeting composers sit at the top of the feed, so adding context happens exactly where you read it.
Check it worked
Pick a contact who has been called by a campaign and read their feed as if you were them. If the sequence looks unreasonable — three calls in a day, a message at 10pm — you have found something to fix in the campaign rather than in the feed.
The parts people get wrong
- Every campaign attempt lands here as a call log entry. A contact rung four times shows four entries, which is what makes "why is this person annoyed" answerable in seconds.
- The feed is per contact. It is not a workspace-wide audit log of every action anyone took.
- AI tasks are labelled as such, so you can tell what the voice agent generated from what a person did.