Every Salesforce release ships a long list of features, but Winter '27 feels different for Marketing Cloud Next. Two Agentforce agents, a Marketing Goals Agent and a Marketing Content Agent, move from keynote demo to a real part of the product. For marketing ops teams the question is no longer whether to pay attention to AI. It is how to get your data, templates and approval process ready so the agents help you instead of creating a mess at speed. This post breaks down what is new, what it means for your team, and what to check before you switch anything on in production.

What Winter '27 adds to Marketing Cloud Next

The release leans on three themes: agent-driven automation, simpler campaign execution, and a more consistent personalization language across channels. Release notes and community write-ups point to the same highlights.

One agent for goals, one for content

The Marketing Goals Agent turns a business objective into a campaign framework: it proposes goals, budgets and audiences, then optimizes activity on its own. The Marketing Content Agent generates email and SMS content that follows your brand guidelines and draws on your asset sources. Neither one replaces your team. They produce proposals that you approve, tweak or reject.

Fewer custom Flows, more native actions

Marketing Completion Actions handle the tasks that follow an interaction, such as an internal notification or an autoresponder email, without a custom-built Flow. The redesigned Campaign Overview pulls planning, execution and performance into a single view, and in some cases you can now edit tracked link destinations after sending.

Handlebars and AMPscript across every channel

This is probably the most useful change day to day. Personalization logic now extends to SMS, WhatsApp, RCS, in-app messages and push, and it also works in landing pages and custom HTML forms. One piece of logic can serve every channel.

The one thing to remember: an agent is only as good as the data and templates behind it. Feed it vague brand rules or poorly defined audiences and it will simply scale your existing mess, faster.

What this means for your team

The marketer's role shifts from hands-on production toward framing and review. Three workstreams are worth starting now.

1. Sharpen your goals and audiences

A goals agent only works if your goals are measurable. For every program, write down one primary KPI, a budget ceiling and the eligible audiences. It doubles as a maturity check: if a human struggles to define it, an agent will too.

2. Clean up and document your content library

The Content Agent relies on your guidelines and assets. Consolidate approved blocks, retire outdated duplicates, and spell out tone of voice, legal footers and phrases you never want to see. Keep consent and unsubscribe language non-negotiable.

3. Set an approval model

Decide who approves what, at which stage, and with how much autonomy. A simple rule works well to start: the agent proposes, a human signs off on anything that reaches customers, and you loosen the reins gradually on low-risk messages.

A practical cross-channel personalization example

Extending Handlebars and AMPscript to conversational channels makes shared personalization logic realistic. Here is an illustrative AMPscript snippet. Adapt it to your data model and test it in your own environment:

%%[
VAR @firstName, @tier, @greeting
SET @firstName = AttributeValue("FirstName")
SET @tier = AttributeValue("LoyaltyTier")

IF NOT EMPTY(@firstName) THEN
  SET @greeting = CONCAT("Hi ", @firstName)
ELSE
  SET @greeting = "Hi there"
ENDIF
]%%
%%=v(@greeting)=%%, your %%=v(@tier)=%% offer is live until Sunday.

The script itself is not the point. The principle is: define a business rule once (a fallback greeting, a loyalty-tier branch), document it, and reuse it in email, SMS or push instead of maintaining three drifting variants.

What to check before you activate

Before turning on any agent in production, validate the following in a sandbox.

• Permissions and roles tied to the agents, so you limit what they can create or send.
• Traceability: who approved which content, and based on which proposal.
• Template compatibility with the cross-channel logic, especially special characters in SMS.
• Deliverability impact, because more content volume must never hurt your sender reputation.

Also keep in mind that exact feature availability depends on your edition, region and rollout schedule. Check the official Salesforce release notes and confirm with your Salesforce contact before you commit to a project plan.

And what about Marketing Cloud Engagement?

If you still run on Marketing Cloud Engagement, Winter '27 includes a guided setup to connect it with Marketing Cloud Next and Data 360. It is not a call to migrate tomorrow, but it is a clear signal that the roadmap investment is going into Next. An audit of your journeys, Data Extensions and existing AMPscript is the best starting point for a realistic timeline.

Key takeaways

Two agents, two jobs. The Marketing Goals Agent frames campaigns, while the Marketing Content Agent produces email and SMS content at scale.

Data comes first. Without measurable goals and a clean content library, agents amplify what is already broken.

Personalization goes cross-channel. Handlebars and AMPscript now apply to SMS, WhatsApp, RCS, in-app and push.

Governance before go-live. Decide who approves what before an agent touches production.

Start preparing now. Want an audit or a readiness plan? Get in touch with CGC-Agency to talk it through.