TrustPager is built to work with agents — think of agents like digital assistants that complete tasks you assign them. Every action your agent takes inside your CRM passes through a scope that you set.
Scopes are the rules — what’s allowed, what needs your approval, and what’s off the table. The system shapes itself around how your business actually runs, so the agent follows your patterns instead of imposing its own.
Watch the walkthrough
Two minutes end to end — set the rules, then watch them play out as the agent sends an email, fires a text, and asks before deleting a duplicate.
Where to find them
From any page in TrustPager:
- Open the Auto tab in the top navigation.
- Click AI Access in the Auto sidebar.
- The Scopes panel is the default view — each resource sits on its own slider.
These settings control what your agents are able to do — and what they can’t do without your approval.
How the sliders work
Each resource (Contacts, Email, SMS, Pipelines, and so on) has a slider that runs left to right. The label on the right tells you what the current position means.
- No Access — the agent can’t touch the resource at all.
- Read — the agent can see the data but can’t change it.
- Read + Write — the agent can read and update, but not delete or send.
- Send with Approval (Email, SMS, Calls) — the agent drafts the message; you review and approve before it goes out.
- Delete with Approval (CRM resources) — the agent proposes the delete; you sign off before it happens.
- Full Access — the agent does it on its own, no approval needed.
Drag the knob to where you want it. Nothing’s official until you hit Save Changes in the top-right.
What this looks like when an agent is working
Email and SMS — review, edit, send
Ask the agent to send an email and the draft comes to you first. You’ll see it in the Approvals queue (shield icon in the top-right of any page). Open it, tweak the subject or body in place, then hit Approve when it reads the way you would have written it yourself.
This isn’t a roadblock — it’s a final pass. You’re the one holding the pen; the agent just gets the page started.
The same flow works for text messages. Prompt, draft, quick tweak, send.
Routine work just flows
Add a contact, update a phone number, move a deal between stages? Those are read/write actions — they land instantly, no approval needed. The big calls — sending messages, deleting records — are the ones that wait for your nod.
That asymmetry is the point: routine work moves fast, important calls keep you in the loop.
Deletes always ask
By default, every delete action in TrustPager sits on Delete with Approval. The agent proposes the delete; you approve or reject in the Approvals queue. The cost of an accidental delete is high; the friction of one extra click is tiny.
Your assistant. Your terms.
Your agent can handle as much or as little as you like — it’s yours to run, whatever way your business needs. Loosen something when you trust it; tighten it when you don’t. There’s no wrong answer, only the one that fits how your team actually operates.
Pair this with: Connect Claude to TrustPager if you haven’t plugged an agent in yet. Then come back here and set your scopes.
