Classifier

The classifier is the behavioral model that produces the routing decision for every send. It runs between your agent.send() call and the actual delivery, reading signal families that capture how the send behaves rather than what it claims to be.

The five signal families

• Volume curve. Ramp-up rate, time-of-day distribution, batch size, inter-send gap. Cold campaigns burst; transactional trickles.
• Recipient sourcing. Have you corresponded with this address before? Is it on your verified contact list? Was it bulk-loaded? New addresses with no prior touch are cold-suspected.
• Engagement curves. Open rate, reply rate, complaint rate, bounce rate, unsubscribe rate. Post-send signals re-classify future sends.
• Content patterns.Cold-script tropes (“I came across”, “noticed you”, “quick question”), merge-tag density, link density, attachment patterns.
• Per-agent baseline. Each agent develops a behavioral signature over its first ~200 sends. Deviation from baseline triggers re-routing.

Transparency

Every send response includes classifier_score(0–1, where higher means more likely to route to Mixed or Outbound) and top_signals— the four highest-weighted features that drove the decision. You can see exactly why a send went where it did.

If you disagree with a routing decision, request manual review from the dashboard. Reviews complete within an hour during business hours and update the per-agent baseline so future sends route correctly.

Read the three-pool post for the full rationale, or the three-pool routing glossary entry for the pool definitions.