Decisioning is the per-user optimization capability in Agent One Teams. Before each journey message goes out, it selects the content, channel, and time most likely to work for that individual user. Your audience receives a message built around them, not one built for everyone.
Decisioning answers three questions for every user: which message variant they should receive, which channel should reach them, and when to send it. It learns from each user's own behavior and keeps adjusting as that behavior changes.
In this article:
Why marketers love Decisioning
One decision per user, not per audience bucket. Instead of sending everyone the same message on the same channel at the same time, Decisioning picks the combination most likely to move each user toward your goal.
Optimization that keeps learning. Decisioning is not a one-time setup. It learns continuously from how each user responds, so its choices stay current as behavior changes.
Control that stays with you. Decisioning can make up to three decisions per send: content, channel, and time. You choose how many of the three it controls and at what granularity. Hand it all three, or hand it one and keep manual control of the rest.
What Decisioning decides
Decision | What it answers | How it decides |
|---|---|---|
Content | Which message variant should this user receive? | You write a base content, and Decisioning produces labeled variants from it. At send time, it matches each user's learned content affinity (tone, value proposition, incentive, urgency, and more) to the best variant. |
Channel | Which channel should reach this user? | Decisioning selects the channel this user is most likely to engage with, from the reachable channels available. |
Time | When should this user be sent to? | Decisioning selects the send time this user is most likely to engage at. |
Content: base content and variants
The content you write inside the channels under the Decision Agent is the base content. Decisioning takes that base content and produces a set of variants from it. Each variant carries content labels such as tone, value proposition, incentive, and urgency. At send time, Decisioning matches the best variant to each user based on what that user has responded to before.
You stay in control of what goes live. Decisioning produces variants from your base content with your brand voice and content guardrails applied. You review and approve the variants before you launch the journey.
Decisioning in action
One offer, many angles
Your team has one offer to send, but your users respond to different things: some to a friendly tone, others to a clear incentive, others to urgency. You write one base content. Decisioning produces labeled variants, and each user receives the variant that matches what they have responded to before.
The right channel at the right time
Your message is ready, and your users are reachable on multiple channels. Set Decisioning to control channel and time, and keep the content under your manual control. Each user receives your message on the channel and at the time they are most likely to engage.
How Decisioning works in a journey
Decisioning is the intelligence layer of Agent One Teams. Inside a journey, it lives in a single element, the Decision Agent. You place the Decision Agent between your audience and your channel sends. As each user reaches it, Decisioning routes them to the best content, channel, and time for them, and the journey sends accordingly.
Use Decisioning for a single message
In Architect, place one Decision Agent before the message you want to optimize.
Write your base content inside the channels under the Decision Agent.
Choose which decisions Decisioning controls (content, channel, time) and at what granularity.
Review and approve the variants produced from your base content.
Launch the journey.
Use Decisioning across several messages
Decisioning works at the level of a single send. It optimizes one message at the point it is about to go out, for the user in front of it. To use it across a flow of more than one message, place a separate Decision Agent before each message in the journey.
Each message is decided on its own. Messages sent one after another are not evaluated together as a sequence. Every send gets its own independent decision based on the user's behavior at that moment.
Limits and good to know
Decisioning does not yet optimize a multi-message sequence as a single connected decision.
Each message in a journey needs its own Decision Agent.
To set expectations, here is what Decisioning is not:
It is not rule-based branching. No fixed "if country = X" conditions sit behind it. It learns and decides per user.
It is not a segment-level optimizer. It makes decisions for the individual user, not for an audience bucket.
It is not a one-time setup. It keeps learning from how each user responds and updates its decisions over time.
Next steps
Decisioning replaces one-size-fits-all sends with a decision made for each user: the right content, on the right channel, at the right time. Ready to see it in action? Place a Decision Agent before your next journey message in Architect, and let Decisioning do the matching.