Author a messaging sequence framework without writing message copy into fields that are meant to hold framework.
A Messaging Sequence is a reusable framework. It records why the sequence exists, which artifacts AI may reference, how many steps it runs, what each step should accomplish, and the overall style. What it does not contain, in any field, is the message. AI writes that later, per account and per contact, when the sequence runs. The detail view reflects the same discipline: it shows the framework, never one account's generated message.
Almost every problem teams hit on this surface traces back to ignoring that split, and it usually happens for a mundane reason. The style guidance box is the largest free-text field available, so somebody writes a nicely-crafted email into it. The campaign then sends near-identical messages to every contact, which is precisely the outcome per-contact generation exists to avoid, and it is hard to diagnose from the output because each individual message reads well.
The purpose field is the instruction AI generates against, so the outcome-versus-activity distinction returns here exactly as it applied to the campaign brief. "Follow up" names a motion and specifies nothing about what the message must achieve. "Re-engage a dormant executive sponsor" names a result, and carries the audience, the situation and the job the message has to do.
The same applies step by step. Set the number of steps, then give each one a single purpose and a message type. Steps run in the order shown. A step trying to both educate and ask for a meeting produces a message that hedges on both, for the same reason an audience spanning two objections produces hedging copy. If you later reduce the step count, the app asks before discarding a step you have already written.
Style attributes are what keep a hundred generated messages recognizably from the same company. Set them against your actual voice rather than a generic idea of professional, and check they agree with the messaging guidance already in your artifact library instead of quietly competing with it.
When drafts come back reading subtly wrong, style is the first place to look, and it is the cheapest thing to fix. Changing a style attribute changes every subsequent generation. Editing an individual message changes one. That relationship is the whole principle behind the next module, arriving here in a smaller form.
The sequence for the example campaign's converters, as it should actually be authored. Purpose: "Move a mid-market RevOps lead who downloaded the evaluation guide from research mode to a scoping conversation." Reference artifacts: the ICP, the Product Overview, the mid-market case study. Steps: three.
Step 1, email — purpose: "Confirm the guide landed and surface the one question the download implies they are asking." Step 2, email — purpose: "Address the migration-risk objection directly, using the case study as evidence." Step 3, email — purpose: "Offer a 30-minute scoping conversation with a named colleague, making the ask specific and small." Style attributes: direct, peer-to-peer, no exclamation marks; guidance box: "Write like an operator who has run this migration, not like a vendor describing it."
Every field is framework. The words each recipient reads will be generated per contact when the sequence runs, referencing their company, their download, their role. Notice the guidance box entry is tone instruction, thirteen words, and could not be mistaken for an email.
Do this in the product
You should be able to answer each of these from memory before opening it. Recalling the answer is what makes it stick; recognizing it when you read it does not.
It contains why the sequence exists, which artifacts AI may reference, the number of steps, what each step should accomplish, and the style. It does not contain the message, which AI writes per account and per contact at run time.
Every contact gets near-identical copy, defeating per-contact generation. It is hard to spot from the output because each individual message reads well.
It records why the sequence exists, which artifacts AI may reference, how many steps it runs, what each step should accomplish, and the overall style. AI writes the message later, per account and per contact.
Each step gets a purpose and a message type, and steps run in order. A step doing two jobs produces a message that does neither.
Paste a finished email into it and you get that email, near-identically, for every contact. Which is the opposite of what per-contact generation is for.
See agents carry the repeatable work of GTM across sales, marketing, customer success, and RevOps. Every action prepared, reviewed, and recorded. Fictional data, real product.
Explore the demo