Deliverability is a function of sender reputation, and sender reputation is tied to the domain you send from. When your AI agents send email from a shared domain pool, they inherit the reputation — good or bad — of every other sender on that pool. For production agent workflows that send business-critical email, you want your agents sending from your own verified domain.
What BYOS Means
AgentMailr's Bring Your Own SMTP (BYOS) feature lets you configure any outbound email provider at the organization level. All emails sent by your agents use your provider and your sender domain. AgentMailr handles the agent logic — inbox management, OTP extraction, threading, webhooks — while your provider handles delivery.
Connecting AWS SES
// Configure AWS SES as outbound provider
await client.smtp.configure({
provider: "SES",
aws_access_key_id: "AKIAIOSFODNN7EXAMPLE",
aws_secret_access_key: "wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY",
aws_region: "us-east-1",
from_domain: "agents.yourcompany.com",
});
// All subsequent sends use SES automatically
await client.inboxes.send("inb_abc123", {
to: "customer@example.com",
subject: "Your order has shipped",
html: "<p>Your order #12345 shipped today.</p>",
});
Why AWS SES at Scale
At production volume, the economics of dedicated sending are dramatic. AWS SES charges $0.10 per 1,000 emails — compared to $1.50-$3.00 per 1,000 for shared pool providers. An agent sending 100,000 emails per month pays $10 on SES vs $150-$300 on a shared pool. The setup investment (domain verification, DKIM configuration, production access approval) pays for itself within weeks.
Supported Providers
- AWS SES — cheapest at scale, $0.10/1k, requires production access approval
- Resend — developer-friendly, free tier, best for getting started
- SendGrid — enterprise features, dedicated IP pools
- Postmark — transactional specialist, high deliverability reputation
- Raw SMTP — any SMTP server, including self-hosted Postfix or Haraka
Fallback Behavior
If your configured SMTP provider is unreachable, AgentMailr falls back to its shared delivery pool automatically — ensuring your agents keep sending even during provider outages. You can disable this fallback if strict provider isolation is required for compliance.