Email automation: flows, triggers, and build-from-prompt

The anatomy of a flow

Every lifecycle automation is built from the same primitives: a trigger (an event or condition that enrolls someone), waits (time delays), filters (conditions that gate the next step), and splits (branches, often for A/B tests or behavioral forks). The art is in mapping these to real customer behavior rather than calendar time.

• Welcome / onboarding — triggered on signup; sets expectations and drives first value.
• Abandoned cart / browse — triggered on commerce events; among the highest-ROI flows.
• Post-purchase — triggered on order; nurtures toward repeat and review.
• Re-engagement / win-back — triggered on inactivity; also protects deliverability by pruning dead contacts.

Why the data model decides everything

A platform can only automate on the data it understands. Klaviyo models profiles plus order data, which is why its commerce flows are so strong. Customer.io models people plus arbitrary custom events, which is why it fits product-led SaaS where the meaningful triggers are in-app actions. Choose the model that matches where your important signals live.

Build-from-prompt: the new authoring model

The newest shift is authoring automations from natural language. Brew lets you describe a sequence — "a three-email welcome flow for new SaaS trial users, with a wait and a branch for users who haven't activated" — and generates the flow on a canvas with trigger, Send, Wait, Filter, and Split nodes that you then refine. It collapses the gap between intent and a working draft.

This doesn't replace the rigor of a well-modeled program — you still need clean events and sensible enrollment logic — but it dramatically shortens the path from idea to first draft, the same way generation shortened creative production.