Name the agent and choose its access
Create an agent in Settings on Starter or Pro. Grant reading activity, recording actions, reading follow-ups, and updating follow-ups separately. Issue a key that can expire or be revoked.
Agent access
Give your AI agent the history behind the work.
Give your agents context from the work already happening across your tools, and keep a visible record of what they do next. Bring an agent you configure; Grondeo supplies the connection and activity history.Illustrative workflow
Source activity and agent workAll activityConnected sourcesAgent work
- ZendeskCustomer issue escalatedEscalated
Ticket #421 · Login issue needs review
- LinearRelated product issue updatedUpdated
Issue #84 · Sign-in retry investigation
- AgentSupport review agent recorded an outcomeRecorded
Reviewed the ticket and related issue
- Follow-upNext step marked resolvedResolved
Follow-up updated by the support review agent
Workflow
From account history to a visible next step
An agent works with the same account activity a person can review, within the permissions granted to its key.
Connect through MCP or REST
Use the six MCP tools or corresponding REST endpoints from an agent you configure. Search recorded activity, open an event, and inspect history related to a source record.
Record the work and next step
The agent can record an action with its outcome and related links, then update a follow-up when its permissions allow. These changes appear alongside other activity in the Timeline.
Example
A follow-up after a customer issue
Illustrative workflow
One question, with the surrounding history
A support agent searches a customer issue, reads the related ticket and product activity, and records the action it took. If it has follow-up write permission, it updates the open follow-up when the work is done.
See how teams review recorded agent work- 01 / ReadSearch activity and related history
Find the ticket, linked issue, and previous account notes.
- 02 / ActWork in the agent’s own tools
The customer supplies and configures the agent that performs the task.
- 03 / WriteRecord an outcome and update a follow-up
The action and follow-up change appear in the account Timeline.
Visibility
See recorded work and agent access
The Timeline shows actions and follow-up changes an agent records. The Agents tab shows the last 30 days of API and MCP access by assigned agent key, operation, channel, time, and result. It does not show tool inputs or conversations, and a key name does not verify which AI client used it.
Setup
Connect an agent you already use
Creating agents requires Starter or Pro. Start in Agent access settings, then use the guide for your connection method.