Agents remove the copy-paste between AI and inbox, but they also need governance: scopes, approval gates, and audit logs.
The metrics that matter#
- Authentication pass rate: share of mail passing SPF, DKIM, and DMARC alignment. Healthy: 99% or higher.
- Spam complaint rate: Gmail enforces at 0.3%; stay under 0.1%.
- Hard bounce rate: under 2%; above 5% triggers platform reviews.
- Inbox placement: above 90% across major providers via seed tests.
- Engagement: click and reply rates by segment; opens are unreliable after Mail Privacy Protection.
Where to read them#
- Google Postmaster Tools for Gmail domain reputation, spam rate, and authentication.
- Microsoft SNDS and JMRP for Outlook.com IP reputation and complaints.
- Your ESP's delivery and bounce reports.
- DMARC aggregate reports for authentication by source.
Improving the numbers#
- Grant the agent read scope first; add draft scope, then send scope only for defined categories.
- Issue per-agent tokens so access can be revoked individually.
- Require approval for messages to new external recipients.
- Log every action with the reasoning the agent produced.
Frequently asked questions#
What is MCP in email?
Model Context Protocol lets AI clients call tools like 'search inbox' or 'draft reply' through a standard interface. MailMaid exposes an MCP server with scoped tokens.