Inbound leads
Capture company, volume, timeline, and contact details while the visitor is still talking.
How it works
Design the table, then let the agent fill it
- 01
Define your tables
Set up fields for leads, tickets, feedback, bookings, or a custom shape your process needs.
- 02
Connect the agent
Give a subagent permission to write — so the main conversation stays focused while records get saved.
- 03
Collect during the chat
The agent asks for email, intent, or issue details, then writes a complete row — not a messy transcript dump.
- 04
Work it as a team
Open the record, assign a person, and track status the same way you would in a lightweight CRM.
Capabilities
What datastores are for
If it should become a row someone acts on, it belongs here.
Support tickets
Open a ticket only when a human is needed — with context already filled in.
Feedback & reviews
Ask follow-ups, classify sentiment, and store structured product feedback.
Booking requests
Collect preferred times and contact info, then queue a request for your calendar owner.
Custom tables
Build the fields your business cares about — not a one-size CRM template.
After the conversation too
Conversation intelligence can write datastore entries once the chat or call ends.
In practice
Why not paste the transcript into a CRM by hand?
Because nobody does that consistently — and when they do, fields are missing.
Complete records
The agent asks for the fields you defined before it saves, so rows are usable on day one.
Same shape every time
Reporting and handoffs work because every lead or ticket looks alike.
Scoped to your org
Records stay with your organization — not mixed into someone else's workspace.
Works with voice
Spoken conversations can write the same tables as chat.
FAQ
Common questions
Is this a full CRM?
It is a structured store your agents can write to and your team can work from. Many teams use it as the intake layer, then sync important rows to the tools they already live in.
Can the agent update an existing record?
Yes. Subagents can create or update rows based on what the visitor said — for example adding notes to an open ticket.
Does conversation intelligence use datastores?
It can. After a conversation ends, a reviewer can save leads, tickets, or other entries automatically.