Marketing Cloud Next is the agentic generation of Salesforce marketing — built on the Agentforce 360 Platform with Data 360 at its core, with agents that can draft a campaign brief, build an audience from plain language and set up a journey in Flow. The most useful thing we can tell you about it first: Salesforce says existing Engagement and Account Engagement customers reach it with no migration.
The most common thing we are asked about Marketing Cloud Next is how big the migration is. The answer, from Salesforce, is that there is not one: Marketing Cloud Engagement and Account Engagement customers can access Marketing Cloud Next capabilities with no migration required.
Next layers new capability alongside the journeys, workflows, content and data that already exist in your org, rather than replacing them. That is a materially different proposition from the re-platforming many partners are quoting for, and it changes what a sensible program looks like.
Four things follow from it.
The journeys, automations, templates, data extensions and engagement programs you have built continue to work. Adopting agentic capability is additive, so the business case does not have to carry the cost of rebuilding what already functions — which is usually the largest line in a migration estimate.
No migration does not mean no work. Which capabilities to enable, in what order, what has to be true in your data first, and who approves what an agent produces are all real decisions. That is a bounded advisory engagement, and it costs a fraction of the migration it replaces.
Growth and Advanced are the first two editions of Marketing Cloud Next on the core platform, and the add-ons extend them. Which you need, and whether an add-on is justified, is a genuine architectural question — it is simply not a migration question.
This is the clearest qualifying question available in this market right now. A proposal to migrate off Engagement or Account Engagement in order to reach Next capability should be accompanied by the Salesforce documentation requiring it. There is a reasonable chance it will not be.
Growth and Advanced are the first two editions of Marketing Cloud Next in the core Salesforce platform, with Data 360 as their foundation. The add-ons extend them. Here is what actually separates them.
| What it is | Best fit | |
|---|---|---|
| Marketing Cloud Growth Edition | Marketing automation built directly on the Agentforce 360 Platform, with Data 360 as its foundation — so marketing works from the same records as sales, service and commerce, with Agentforce available. | Growing businesses standardizing on the core Salesforce platform, who want marketing on the same data model as everyone else. |
| Marketing Cloud Advanced Edition | Growth plus advanced predictive AI, path experimentation, Unified Conversations for SMS and WhatsApp, and higher Flow, messaging and AI credit allowances. | Organizations that have outgrown Growth — particularly where two-way conversational messaging or experimentation matters. |
| Salesforce Personalization | A Next add-on: real-time interaction management on web and app. | Digital properties with meaningful anonymous traffic. See Personalization for what it needs first. |
| Marketing Intelligence | A Next add-on: cross-channel marketing data visualization and reporting. | Multi-channel spend that has to be reported as one number. See reporting and analytics. |
| Loyalty Management | A Next add-on: loyalty program management alongside the marketing platform. | Businesses running a loyalty or rewards program that should share the customer record. |
| Marketing Cloud Engagement | The established B2C suite. Still live, still sold, and able to reach Next capabilities without migrating. | High-volume consumer messaging with complex data and multi-brand estates. See the practice page. |
| Marketing Cloud Account Engagement | The B2B product, formerly Pardot. Customers on Growth, Plus, Advanced and Premium editions can access Next capabilities in supported regions. | B2B demand generation with a sales-led funnel. See Account Engagement. |
Capabilities verified against Salesforce documentation Last reviewed 23 September 2026
Described as steps a marketer currently performs, because that is the useful comparison. Salesforce's own framing is that agentic marketing lets a marketer define the strategy while agents handle the execution, and that campaigns can be deployed in hours rather than weeks.
You describe a goal in natural language and Agentforce produces a campaign brief, then works through the execution: creating the audience, drafting email and SMS content, and setting up the customer journey in Flow. The marketer collaborates and approves rather than assembles.
Unified Conversations for SMS and WhatsApp, available in Advanced Edition and Next, turns one-way promotion into an exchange the customer can answer. This is a bigger operational change than it appears — a channel that replies needs someone to reply to it.
Rather than setting a campaign live and reviewing it afterwards, the agentic model monitors performance against the goal and adapts within the guidelines you set. The guidelines are the work — an agent optimizing toward an unstated objective will optimize toward the wrong one.
The quality of anything an agent produces depends on the data it is grounded in. Data 360 is the foundation of Growth and Advanced, which is why data readiness is the first item in our adoption practice rather than an afterthought.
Before any conversation about adopting new capability, it is worth checking what is switched on in the org you already have. Einstein features in Marketing Cloud Engagement are frequently unused, and just as frequently duplicated by custom code written before they existed.
They are not switch-on magic, and the reason is always the same: they learn from your engagement history, so they need one. Stating the prerequisite is not a caveat — it is the thing that determines whether enabling them now is worth doing.
Retiring custom code that duplicates a native feature is one of the cheaper wins an audit produces.
Predicts how likely a contact is to open, click, stay subscribed or convert, using machine learning over your own engagement history. Scores can drive audiences, persona segments, journey splits and content personalization. Salesforce documents a requirement of at least 1,000 events in the 90 days before activation; scores are typically available 24 to 48 hours after implementation, with the model maturing over 90 days.
Chooses the best moment within the next 24 hours to send to each individual contact, using 90 days of email or push engagement data. Available in Journey Builder as well as for standalone sends. This is the feature most often reimplemented as custom code by teams who did not know it existed.
Selects which content to show each customer when they engage with a message, rather than requiring a rule for every combination. Useful where the number of content permutations has outgrown what dynamic content blocks can sensibly express.
All of these learn from engagement history, which means a newly migrated org, a newly warmed sending infrastructure, or a program that has just changed audience does not have the history yet. Enabling them and expecting immediate results is a timing problem rather than a product one — and it is the commonest disappointment we hear about.
The technology question — can an agent draft this? — is largely settled. The unresolved questions are organizational, and they are the ones that determine whether the capability gets used or quietly abandoned after the first uncomfortable draft.
None of these are AI-specific in principle. All four become urgent the moment execution stops being manual.
What may go out without a human — A spectrum, not a switch: agent drafts with human approval, agent executes within preset bounds, agent adapts a live campaign autonomously. Different message types warrant different positions, and deciding that in advance is considerably easier than deciding it during an incident.
What the brand will not say — Claims, comparisons, guarantees, regulated language, tone. An agent generating content needs the boundaries written down — which most organizations have never had to do explicitly, because until now the boundaries lived in the judgement of the people writing.
What data an agent may reason over — Grounding in more data makes output better and makes the governance question sharper. Sensitive attributes, inferred characteristics and data collected for another purpose all need an explicit position before an agent has access to them.
How you would know it went wrong — Audit trail of what the agent changed and why, monitoring of what is being produced, and a defined stop — who halts it and how quickly. Autonomous execution without an obvious brake is the reason most cautious organizations stall at pilot.
Agentic capability amplifies what is already true of your org. Good data and clear governance get better leverage; poor data and unclear ownership get faster, more confident versions of existing problems.
If more than two of these are uncertain, the honest first engagement is readiness rather than enablement.
The grounding layer. Data 360 is the foundation of Growth and Advanced, and of anything an agent reasons over.
Edition selection, architecture and how a larger engagement runs.
What MCAE customers get from Marketing Cloud Next, and on which editions.
Unified Conversations, and the consent work that has to precede a channel that replies.
Establishing what Einstein capability you already own and what custom code duplicates it.
A worked example of the architecture-first approach on a real client engagement.
No. Salesforce's published position is that Marketing Cloud Engagement and Account Engagement customers can access Marketing Cloud Next capabilities with no migration — Next layers new capability alongside the journeys, workflows, content and data already in your org rather than replacing them.
What is worth planning is adoption: which capabilities to enable, in what order, and what has to be true in your data first. That is a readiness engagement. If a partner is quoting you a migration, ask them to point at the documentation that requires one.
Einstein features in Marketing Cloud Engagement are predictive models that answer specific questions — how likely is this contact to click, when should we send to them, which content should they see. They inform decisions a marketer or a journey then acts on.
Agentforce is an execution layer. Rather than answering a question, an agent carries out a sequence of work: generating a brief, building an audience, drafting content, setting up a journey. Different capability, different governance requirement — and the two coexist rather than replacing one another.
You describe the goal in natural language and it produces a campaign brief, then works through the execution: creating the audience segment, drafting email and SMS content, setting up the customer journey in Flow, and providing a campaign summary. Salesforce's framing is that a campaign can be built and deployed in hours rather than weeks.
Each step is reviewable. In practice the value is less about speed on any one campaign and more about capacity — the campaigns that never got built because nobody had time now become viable.
That is the right question and it is the first thing we assess. Agentic capability amplifies what is already true of your org: good data and clear governance get real leverage, while poor data produces faster and more confident versions of existing problems.
The specific things that matter are identity — can you tell customers apart reliably — consent state, and whether the attributes an agent would reason over are current. Data 360 is the foundation of Growth and Advanced, which is why data readiness leads our adoption sequence rather than following it.
Check what you already own. Einstein Engagement Scoring, Send Time Optimization and Content Selection are available in Engagement and are frequently unused — and frequently duplicated by custom code written before they existed. An audit establishes both.
The prerequisite is engagement history: Salesforce documents a threshold of at least 1,000 events in the preceding 90 days for Engagement Scoring, and Send Time Optimization uses 90 days of engagement data. A recently migrated org needs to accumulate that before any of it will perform.
By writing down what the brand will not say — claims, comparisons, guarantees, regulated language, tone — which most organizations have never had to do explicitly, because those boundaries have always lived in the judgement of the people writing.
Then by deciding what may go out without a human, per message type, and by keeping an audit trail and an obvious stop. The governance work is the adoption work; the technology is the straightforward part.
Tell us which edition you are on and what state your customer data is in. That is usually enough to say which capabilities are worth enabling now and which should wait for the data to catch up.
Serving: US | Canada | UK | UAE | Australia | New Zealand
