Salesforce Winter ’27: The AI and Automation Updates Businesses Need to Know

Salesforce Winter ’27 update is here! 🚀 Let’s unpack the Salesforce updates that actually matter.

Every Salesforce release comes with a long list of new features, enhancements, changes, and release updates. The difficult part isn't finding that information. It is figuring out what actually matters to your business and your Salesforce org.

If you're using Salesforce today, you probably don't need another article that lists dozens of new features. You need to know:

• Which Winter ’27 updates could affect your existing Salesforce setup?

• Where does Agentforce fit into your current workflows?

• What role does Data 360 play in Salesforce's AI strategy?

• Which Flow and automation changes should admins pay attention to?

• What should developers test?

• And, perhaps most importantly, where should your business actually start?

The bigger story behind Salesforce Winter ’27 is the closer connection between AI, customer data, automation, and CRM workflows. Agentforce provides intelligence and action, Data 360 provides customer context, while Flow, Apex, APIs, and the Salesforce platform provide the tools needed to execute business processes.

The real value isn't one individual feature. It's how Salesforce is bringing AI, data, and automation together inside everyday CRM workflows.

Salesforce Winter ’27 at a Glance

Here’s a quick look at the Salesforce Winter ’27 areas that matter most for admins, developers, and business teams.

These capabilities are increasingly connected, bringing AI, data, automation, and CRM workflows closer together.

1. Agentforce: The Biggest Winter ’27 AI Story

One of the biggest stories in Winter ’27 is Agentforce. To understand why it matters, it helps to look at how Salesforce's AI experience has evolved.

Before: AI as an Assistant

Earlier Salesforce AI capabilities were largely focused on helping users:

• Generate content

• Summarize information

• Find insights

• Retrieve customer information

• Get assistance inside CRM workflows

For example, an AI assistant might help a salesperson understand what happened with an account or summarize an opportunity.

Now: AI That Can Participate in the Workflow

The progression is straightforward:

AI assistant → AI agent → AI agent that can take action

An agent is not simply answering a question. It can potentially use Salesforce data, available actions, automation, and connected systems to help move a business process forward.

For example, imagine a service representative receives a customer request. The traditional process might require the representative to:

1. Find the customer

2. Review the account

3. Search previous cases

4. Look for relevant knowledge

5. Decide what to do

6. Update Salesforce

7. Respond to the customer

With an agent-based workflow, AI can potentially assist with several of these steps.

The real value is less manual work + faster decisions + better use of CRM data.

An assistant might tell a sales representative what happened with an account, while an agent can potentially take the next step using the appropriate Salesforce data, actions, and automation.

What is changing with Agentforce?

Rather than focusing on every individual Agentforce release note, businesses should look at a few broader areas:

• Agent actions

• Agent customization

• Agent analytics and observability

• Agent usage

• Agentforce and Flow

• Agentforce and Salesforce data

• Connections to external systems and tools

AI is becoming more connected to the systems where work actually happens. An agent could potentially assist with tasks such as searching Salesforce, checking information, deciding what to do, and updating records.

The benefit is less manual work + faster decisions + better use of existing CRM data.

What does Agentforce mean for businesses?

The first question shouldn't be:

"Where can we use AI?"

It should be:

"Which business processes are consuming too much employee time?"

Look for repetitive activities such as:

• Answering common customer questions

• Summarizing customer information

• Preparing follow-ups

• Finding relevant records

• Supporting service representatives

• Assisting sales teams

• Guiding employees through processes

• Taking actions based on business rules

This gives you a much better starting point than simply turning on an AI feature because it is new.

What should Salesforce admins and developers consider?

Agentforce introduces another layer of responsibility. Admins and developers need to think about:

• Permissions

• Data access

• Agent actions

• Flow

• Integrations

• Testing

• Monitoring

• Human approval

An AI agent is only useful when it has access to the right information and the right actions.

Giving an agent access to everything in Salesforce simply because it makes implementation easier is not a sound approach.

Salesforce's Winter ’27 documentation already includes new Agentforce capabilities across areas such as external API and MCP connectivity, agent configuration, and service experiences.

What does this mean for Salesforce users?

For everyday users, the biggest change may be how work gets done inside Salesforce.

Instead of switching between Salesforce records, documents, knowledge articles, reports, and external systems, users may increasingly interact with AI directly within their workflow.

2. Marketing Cloud Next: AI Moves Deeper Into Marketing

Marketing Cloud has traditionally helped marketing teams manage campaigns, audiences, content, customer journeys, and automation. Marketers still needed to define the campaign, select the audience, create the content, build the journey, and monitor the results.

Winter ’27 pushes this workflow further with AI.

Instead of thinking about AI as just a tool for generating marketing content, Salesforce is moving toward a model where AI can assist across more stages of the marketing process.

Before:

Marketing goal → Marketer plans campaign → Select audience → Create content → Build journey → Launch → Analyze results

That is more meaningful than simply saying that marketers now have AI tools.

The real question is whether AI can help marketing teams move faster from an idea to an actual customer experience.

Winter ’27 brings attention to areas including:

• Marketing-focused Agentforce

• AI-assisted campaign planning

• Marketing Goals Agent

• Personalization

• Journey decisioning

• Customer data and marketing

For marketers, this could change how campaigns are planned and optimized.

Instead of starting with a blank campaign and manually working through every step, AI can increasingly assist teams with planning, audience understanding, content, and customer journeys.

But there is a catch.

AI-generated marketing is only as useful as the customer data behind it.

If your customer profiles are incomplete, duplicated, outdated, or spread across disconnected systems, adding more AI won't magically solve the problem.

How could the marketing process change?

Earlier workflow:

A marketer would typically need to review customer data, define a target audience, create campaign content, configure the journey, set decision rules, launch the campaign, and then analyze performance.

AI-assisted workflow:

Marketing Cloud Next can help teams move from a business goal to campaign execution with less manual work. For example, a marketer looking to increase repeat purchases could use AI to assist with campaign planning, audience understanding, personalization, and customer journey decisions. In contrast, the marketer remains responsible for the overall strategy.

But there is an important catch: AI needs good customer data.

This is where Marketing Cloud Next connects with the wider Salesforce platform.

AI can help marketers work faster, but it cannot fix poor customer data by itself.

If customer profiles are:

• Incomplete

• Duplicated

• Outdated

• Stored across disconnected systems

• Missing important customer interactions

Then AI may have limited context when helping marketers make decisions.

Salesforce's broader Winter ’27 story connects marketing more closely with customer data and Data 360 for exactly this reason: better-connected data creates better customer context, which can make AI more useful.

What should businesses look at?

Before adopting every new AI marketing capability, look at:

• Customer data quality

• Segmentation

• Consent and permissions

• Customer profiles

• Integrations

• Existing journeys

• Marketing automation

The better connected your data is, the more useful AI becomes.

3. Data 360: Why Data Matters More in the AI Era

AI needs context.

Consider a simple question from a customer:

“Why hasn't my order arrived yet?”

If an AI system can only see a basic customer record, it has limited context.

But what if it can access relevant information about:

• The customer

• Their order

• Shipment status

• Previous service interactions

• Account history

• Relevant business information

Now the AI has a much better foundation for providing a useful response or supporting the next action.

This is why Data 360 becomes increasingly important.

From Scattered Data to Customer Context

Many businesses have customer information spread across:

• Salesforce

• ERP systems

• Marketing platforms

• Spreadsheets

• Support applications

• Commerce systems

• Other external tools

The challenge is not simply having more data. The challenge is creating useful, connected, trustworthy data.

A simplified way to think about the relationship is:

CRM data + external data
↓
Unified customer context
↓
AI / Agentforce
↓
Personalization + decisions + actions

The point isn't simply to have more data. It is to have useful, connected, trustworthy data.

Your source highlights Data 360's role around unified customer context, data quality, identity resolution, AI, and data activation.

Why does this matter for Agentforce?

Imagine a customer asks:

“Why hasn't my order arrived yet?”

An agent that can only see a CRM contact record has limited context.

An agent that can access the relevant customer, order, shipment, service, and interaction information has a much better chance of providing a useful response or taking the appropriate action.

That is why data quality becomes an AI issue.

Poor data doesn't just create bad reports anymore. It can also lead to poor AI context.

What should admins and developers review?

If you're considering Data 360 or Agentforce, look closely at:

• Your Salesforce data model

• Data quality

• Duplicate records

• Identity resolution

• Integrations

• Data access

• Governance

• External data sources

A simple rule is worth remembering:

Better data creates better customer context. Better customer context creates more useful AI.

4. Flow and Automation: AI Doesn't Replace Automation

This is one of the areas where businesses need to avoid getting carried away with AI.

Not every process needs an AI agent. Some processes are predictable and should remain predictable.

For example:

If an opportunity reaches a certain stage → update a field → notify a user.

That's a good automation candidate. You don't need an AI agent to make that decision.

This is where the distinction between traditional automation and agentic automation becomes important.

Traditional automation:

If X happens → do Y

Agentic automation

Understand the situation → determine what action makes sense → execute

Both have a place. Traditional automation works well for predictable processes, while agents can be a better fit for processes that require reasoning and decision-making.

What Flow updates should admins watch?

The Winter ’27 discussion includes areas such as:

• Flow Test Mode

• Flow Tags

• Mass list actions

• Screen Flow improvements

• Flow Builder experience

• Split logic

• Agent-ready Flow capabilities

Salesforce's release documentation and preview information continue to show Flow evolving alongside its broader AI and automation strategy.

How Can Flow and Agentforce Work Together?

Organizations shouldn't think of Flow and Agentforce as competing technologies. They can work together.

Agentforce determines what should happen
↓
Flow executes a defined business process
↓
Salesforce updates the relevant records
↓
The user receives the result

That combination can be much more practical than trying to make an AI agent responsible for every part of a process.

5. Agentforce for Sales and Service

Sales and service teams are two areas where AI can have a direct impact on everyday work.

Rather than treating Sales and Service as completely separate AI stories, it makes sense to look at the common theme:

Salesforce is bringing AI closer to the actual CRM workflow.

Agentforce for Sales

Sales teams often dedicate a significant amount of time to tasks that are not directly related to selling.

Examples include:

• Reviewing account information

• Preparing follow-ups

• Summarizing opportunities

• Finding customer history

• Updating records

• Preparing for meetings

• Tracking deal progress

AI assistance can reduce some of that administrative workload.

The goal isn't to replace the salesperson. It is to give the salesperson more time for conversations, relationships, and decisions.

Agentforce for Service

Service teams face a similar challenge.

A service representative may need to search through customer records, knowledge articles, previous cases, and other information before responding to a customer.

AI can help bring that information together and assist with the next step.

Winter ’27 continues to expand Salesforce's AI-driven service capabilities, including areas such as messaging and Agentforce contact-center experiences.

The important business question is:

How much time can your service team spend helping customers instead of searching for information?

6. What Salesforce Developers Need to Know

Developers should review platform changes before upgrading production environments.

Apex

Developers should review Apex-related changes that could affect:

• Existing code

• Testing

• Security

• Data access

• Performance

• Custom business logic

Focus on anything that could change how your existing implementation behaves.

APIs and Integrations

This is particularly important for businesses with Salesforce connected to ERP, finance, marketing, customer portals, data platforms, or other external applications.

Salesforce's Winter ’27 API documentation includes new and changed API capabilities, including access to additional objects and metadata through API version 68.0.

Before production upgrades, developers should review:

• API versions

• Authentication

• External integrations

• Middleware

• Custom API calls

• Data synchronization

• Error handling

An integration that works today should not simply be assumed to work exactly the same way after a major release.

Lightning and Platform

Developers should also review relevant changes involving:

• Lightning Web Components

• Metadata

• GraphQL

• Platform capabilities

• Developer tooling

• AI-assisted development

Their focus is expanding toward architecture, integrations, governance, APIs, automation, AI actions, and reliable business logic.

Release Updates: What Existing Salesforce Customers Need to Test

There is a major difference between a new feature and a release update.

That distinction matters.

If your Salesforce org has been running for several years, you probably have:

• Flows

• Apex

• Integrations

• Custom objects

• Custom permissions

• Authentication configurations

• Automations

• Third-party applications

• Custom Lightning components

A major release can touch some of these areas.

What should admins test?

At minimum, review:

• Critical Flows

• Apex

• Integrations

• Security settings

• Authentication

• Existing automations

• Customizations

• Deprecated functionality

• Retired functionality

Salesforce's release documentation is actively updated as features and release timelines change, so teams should review the current release notes rather than relying on an old checklist.

What Does Winter ’27 Mean for Businesses?

1. Less Manual Work

AI and automation can help reduce repetitive activities.

The opportunity is particularly strong where employees spend large amounts of time:

• Searching

• Copying information

• Summarizing

• Updating records

• Responding to repetitive requests

• Moving information between systems

2. Better Customer Context

Data 360 and connected data can give both employees and AI a more complete picture of the customer.

That can improve:

• Personalization

• Service

• Sales conversations

• Decision-making

• Customer journeys

3. More Intelligent Automation

Businesses don't have to choose between automation and AI. They can use both.

• Flow can handle predictable processes.

• Agentforce can assist with situations that require more reasoning.

• Apex and APIs can extend the platform when standard capabilities aren't enough.

4. Greater Need for Governance

As AI gains access to more customer and business data, governance becomes increasingly important.

That means paying attention to:

• Data access

• Permissions

• Security

• Monitoring

• Testing

• Human approval

• Auditability

AI adoption without governance can create more problems than it solves.

What Should Salesforce Admins and Developers Do Before Winter ’27?

You don't need to overhaul your Salesforce org just because a new release is available. Instead, take a structured approach.

Step 1: Review Release Updates

Start with the official Winter ’27 release notes and identify changes relevant to your edition, products, integrations, customizations, and retiring functionality.

Step 2: Audit Your Automation

Review your most important Flows and Apex processes.

Ask:

• What processes are business-critical?

• Which automations have the most dependencies?

• Which ones haven't been tested recently?

Step 3: Review Your Data

If you're considering Agentforce, review your data quality first, including completeness, accuracy, duplicates, relationships, access, and external data dependencies.

Step 4: Identify AI Opportunities

Start with the process, not the technology. Identify two or three repetitive activities where AI could realistically save time, and make sure agents have only the permissions they need.

Step 5: Review Agent Permissions

Make sure agents have access to what they actually need—and not more.

Permissions should be part of the design, not something checked at the end.

Step 6: Test in a Sandbox

Salesforce's Winter ’27 sandbox preview window begins August 28, 2026, with preview sandboxes being upgraded ahead of production/non-preview instances.

Use the preview period to test important workflows and integrations.

Step 7: Check Integrations

Review external dependencies, APIs, middleware, authentication, and data synchronization.

Step 8: Check Retirements

Make sure your organization isn't depending on functionality that Salesforce is retiring or changing.

Should Your Business Adopt These Winter ’27 AI Features?

Not every company needs every new Salesforce capability. A simple decision framework can help.

This is a much better way to approach Winter ’27 than asking:

"Which new features should we turn on?"

Instead ask:

"Which business problem are we trying to solve?"

Start with the business problem, then choose the Salesforce capability that best fits the process.

What Should Businesses Prepare for Beyond Winter ’27?

Salesforce is moving AI deeper into CRM processes, so businesses should rethink how they approach automation.

Instead of asking only:

"How do we automate this?"

You may increasingly need to ask:

"Should this be automated with Flow, assisted by AI, handled by an agent, or kept with a human?"

That is a much more useful question.

A mature Salesforce environment may eventually use all four:

• Automation for predictable work

• AI assistance for employee productivity

• Agents for tasks requiring reasoning and action

• Humans for judgment, exceptions, and high-risk decisions

The goal isn't to use more AI. It's to use the right level of intelligence for each process.

Final Thoughts

Salesforce Winter ’27 highlights a broader shift toward connected AI, customer data, automation, and CRM workflows. Agentforce, Data 360, Flow, Apex, and APIs each serve a different purpose, giving businesses more options to improve how work gets done inside Salesforce.

The key is not to adopt every new capability. Start by identifying repetitive tasks, disconnected data, inefficient workflows, or processes that require too much manual effort. Then determine whether Flow, Agentforce, Data 360, custom development, or a combination of these capabilities is the right solution.

If you're planning to prepare your Salesforce org for Winter ’27 or want to modernize your CRM environment, MV Clouds can help with Salesforce implementation, customization, automation, integrations, and AI-ready solutions. Talk to our Salesforce experts to evaluate your current setup and identify practical opportunities for improvement.

Frequently Asked Questions

1. What is Salesforce Winter ’27?

Salesforce's Winter '27 platform release introduces enhancements across a wide range of Salesforce platforms, including Agentforce, Data 360, Flow, Sales, Service, Marketing, APIs, development, security, and platform capabilities. Currently, the release is a preview, so businesses should check the official release notes for individual features and edition requirements.

2. What are the biggest Salesforce Winter ’27 AI updates?

The biggest story is Agentforce's continued expansion and its connection with Salesforce data, automation, and business actions. Rather than using AI only to generate answers or content, Salesforce is increasingly positioning agents as part of actual CRM workflows. This is where they can use the data, tools, and actions to assist with work.

3. What is new in Agentforce in Winter ’27?

Winter ’27 includes developments across Agentforce configuration, actions, service experiences, integrations, and agent capabilities. The most important thing for businesses is not one individual feature but the broader move toward AI agents working with Salesforce data and business processes. Specific capabilities can vary by product and availability stage.

4. What is new in Marketing Cloud Next?

Winter ’27 continues the integration of AI into marketing processes, including campaign planning, personalization, customer journeys, and marketing-focused Agentforce capabilities. The larger opportunity is helping marketing teams move from customer data and marketing goals to campaigns and personalized experiences more efficiently.

5. How does Data 360 support Salesforce AI?

Data 360 helps bring customer information from different sources together so businesses can create a more complete customer context. That context becomes particularly important when AI or Agentforce is expected to provide useful answers or take action. Poor-quality or fragmented data can limit the value of AI.

6. What are the important Winter ’27 Flow updates?

Flow remains a key part of Salesforce automation, with Winter ’27 changes covering areas such as testing, tags, list actions, screen flows, builder improvements, and logic capabilities. The key takeaway is that Flow continues to have a central role even as Salesforce expands Agentforce. Not every business process needs an AI agent.

7. What should Salesforce admins do before Winter ’27?

Admins should review relevant release updates, test critical flows and automations, evaluate integrations, review security and permissions, and verify data quality in a preview sandbox. Before upgrading production, Salesforce suggests evaluating the Winter '27 changes in sandbox preview environments.

8. Do Salesforce developers need to change for Winter ’27?

Not necessarily, but developers should review changes relevant to their org. Apex, APIs, Lightning components, metadata, integrations, authentication, and custom platform functionality should be tested where applicable. The goal isn't to change code simply because a new release exists, but to identify anything that could affect existing functionality.