Insights · AI

AI in Marketing Cloud, without the hype.

Einstein already answers three practical questions inside Journey Builder: who, when and how often. Here is how each one works, and where the new agents fit.

  • AI
  • 7 min read

“AI in marketing” usually means a demo. Inside Marketing Cloud Engagement it already means three features that make decisions in production journeys every day. Used together, they replace the fixed rules most teams still maintain by hand.

WhoEngagement ScoringPredicts how likely each contact is to open, click, stay subscribed or convert, and groups them into personas.
WhenSend Time OptimizationLearns when each contact tends to engage and delays the send to that hour, per person.
How oftenEngagement FrequencyTells you whether a contact is getting too few or too many emails.
The three Einstein questions. Each one is available as a split or activity in Journey Builder.

Who: Engagement Scoring

Einstein Engagement Scoring looks at each contact’s history and predicts the likelihood of the next email being opened or clicked, of the person staying subscribed and of a conversion. It then groups contacts into personas, such as loyalists who engage often and contacts who have gone quiet, so a journey can treat them differently.

When: Send Time Optimization

Instead of sending everyone at 9:00, Send Time Optimization holds each message until the hour when that specific contact usually engages. In a journey it is a single activity placed before the email. It works best on messages that are not time-critical.

How often: Engagement Frequency

Einstein Engagement Frequency reads the last 28 days of engagement with marketing emails, excluding transactional ones, and places each contact in one of four states. In Journey Builder, a Frequency Split routes them accordingly.

UndersaturatedCan receive more
On targetKeep the pace
Almost saturatedSlow down
SaturatedPause or suppress
Einstein Engagement Frequency states in Marketing Cloud, based on each contact’s last 28 days of engagement.

This replaces the classic hard-coded rule of “no more than three emails a week”, which is too much for some people and too little for others.

Where the agents fit

In June 2025 Salesforce announced Marketing Cloud Next, an agent-based marketing product built on the Salesforce Platform, with Agentforce agents that assemble audiences, draft email and SMS content and build journeys. In October 2025 Data Cloud was renamed Data 360 and became part of Agentforce 360.

Our view is pragmatic: agents are useful when the foundations are clean. An agent that builds a segment on duplicated contacts or unclear consent will do the wrong thing faster. That is why every AI project we run starts from the data model.

Where to start

If you are on Marketing Cloud Engagement today, the fastest win is usually Send Time Optimization plus a Frequency Split on your highest-volume journey. Both are measurable within a month.

Key takeaways

  1. Engagement Scoring decides who, Send Time Optimization decides when, Engagement Frequency decides how often.
  2. All three plug into Journey Builder as splits or activities.
  3. Agents need clean data and clear consent to be useful.
  1. Sources: Salesforce, Marketing Cloud Next announcement (June 2025) · Salesforce Ben, Data Cloud renamed Data 360 · Salesforce Ben, Einstein Engagement Frequency

Keep reading

MethodOne signal path, from your data to every channelMeasurementWhy we split on clicks, not opens

Want this inside your instance? Let’s build it together.

Tell us where you are and we’ll tell you what it takes to get there.

Talk to us
Nmc Bubble