Top 10 Salesforce Data 360 Updates Businesses Need to Know

Enterprise data is becoming more valuable and more difficult to manage. Companies today are collecting customer information from CRMs, warehouses, documents, third-party platforms, and AI systems, but turning that scattered data into reliable business insights remains a major challenge.

Salesforce Winter ’27 Data 360 updates address this challenge by focusing on three critical areas: making enterprise data smarter, keeping it secure, and making it easier to use for AI-driven workflows. Instead of simply connecting more data sources, Salesforce is improving how organizations analyze patterns, predict outcomes, control access, and activate information across their business.

From advanced regression models and unsupervised clustering to Snowflake zero-copy connectivity, granular governance, and AI-assisted Document AI configuration, these updates show Salesforce’s continued shift toward building a stronger foundation for Agentforce, automation, and enterprise AI.

1. Predict Numeric Outcomes with Advanced Regression Loss Functions

One of the most useful AI updates is the addition of Gamma and Quantile regression loss functions in Model Builder.

Traditional regression is generally used to predict an average outcome. But many real-world business datasets are skewed. Customer spending, insurance claims, revenue, and delivery times can have a small number of unusually large values.

With the new options, businesses can choose a loss function based on the prediction they actually need.

• Squared Error: Predict the average expected value.

• Gamma: Best for positive, right-skewed outcomes such as customer spend, claims, or revenue.

• Quantile: Predict a specific percentile, such as the 90th percentile of delivery time.

Why it matters

Instead of asking only:

“What is the average expected outcome?”

Businesses can build models around questions such as:

“What will the high-end customer spend look like?”

or:

“What delivery time should we expect for 90% of orders?”

This makes predictive analytics more closely aligned with specific business decisions.

2. Dynamically Group Records with Unsupervised Clustering

Data 360 also adds unsupervised clustering, allowing businesses to discover natural groups and patterns within their data without building and maintaining a reusable predictive model.

The capability can help organizations:

• Identify similar customer or product groups

• Detect unusual records or patterns

• Explore large datasets

• Combine information from multiple fields

• Create clusters that can be used in downstream predictive models

Why it matters

Businesses don't always know the segments they should be looking for before analyzing their data.

Clustering allows the data itself to reveal patterns.

For example, an organization could discover that its customers naturally fall into several behavioral groups based on purchasing frequency, spend, engagement, and product usage.

That information can then support segmentation, forecasting, personalization, and other AI use cases.

3. Enhance Data Security with Granular Data Graph Governance

As organizations bring more information into Data 360, controlling access to that information becomes increasingly important.

Winter ’27 introduces granular governance for Data Graphs. Instead of relying on an all-or-nothing access model, Data 360 can apply governance policies at a more detailed level.

The platform can:

• Prune restricted fields

• Mask data

• Exclude unauthorized child data model objects

Why it matters

This is particularly important for organizations using Data 360 as a foundation for AI.

An AI agent needs access to relevant business context, but it shouldn't automatically have access to every piece of connected data.

To balance access to AI with data security and compliance, organizations need granular governance.

4. Connect Snowflake Data Without Duplicating It

The new Snowflake Zero-Copy V2 Connector expands Data 360's data-sharing capabilities.

Organizations can connect Data 360 with Snowflake and query data as a mounted Snowflake database without duplicating that data into Data 360.

The update also introduces OIDC authentication and support for acting on change data through Snowflake streams and tasks.

Why it matters

For enterprises already using Snowflake, zero-copy connectivity can help reduce unnecessary data movement and duplication.

Instead of creating another copy of large datasets, organizations can work with data where it already exists.

This can be particularly valuable for companies managing large enterprise data warehouses and AI workloads.

5. Run Custom Python Transformations on Zero-Copy Data

Data 360 is also making zero-copy data more flexible by allowing code extensions to run custom Python batch transformations on data backed by zero-copy connectors.

This means organizations can apply Python-based transformations without first copying the external data into Data 360.

Why it matters

This is useful for technical teams that need custom data-processing logic that isn't available through standard transformation tools.

For example, developers could use Python to perform specialized transformations on external datasets while maintaining a zero-copy architecture.

The result is greater flexibility without necessarily introducing another data-copy pipeline.

6. Bring Microsoft OneNote Content into Data 360

The Microsoft OneNote Unstructured Connector is now generally available.

It allows organizations to ingest notebook pages and embedded attachments from SharePoint team notebooks into Data 360.

This is important because valuable business information isn't always stored in structured CRM fields.

It can exist in:

• Meeting notes

• Team notebooks

• Documents

• Attachments

• Other unstructured content

Why does it matter?

Bringing this information into Data 360 can make it available for AI and search experiences, giving Salesforce applications a richer source of business context.

7. Configure Document AI Through Natural-Language Interaction

Another notable update is the Data 360 MCP Server for Document AI.

Organizations can connect an AI agent, such as Claude or Cursor, to the Data 360 MCP server and use natural-language interaction to help:

• Generate document schemas

• Configure Document AI

• Test extraction

• Verify extraction accuracy

• Publish configurations

Why does it matter?

A document AI configuration traditionally involves a number of technical steps.

An MCP-based workflow can make that process more conversational and potentially reduce the time it takes to create and test document-processing workflows.

For development teams, this represents a major shift toward AI-assisted Salesforce configuration and development.

8. Data 360 Engagement Timeline

Salesforce is introducing the Data 360 Engagement Timeline as a new way to view and understand customer engagement within Data 360. It will replace the existing Data Cloud Profile Engagements Widget, giving teams a more unified view of customer interactions.

Why it matters

A unified engagement timeline can make it easier for teams to understand customer interactions and behavior in context rather than looking at isolated activities.

This can support better customer analysis, personalization, and AI-driven experiences.

9. More Flexible Copy Field Enrichments

Winter ’27 also brings several improvements to Copy Field Enrichments, including:

• Expanded field-type compatibility

• Matching on more than the primary key

• Support for picklist-dependent fields during transformation and integration

Why it matters

Data enrichment often becomes complicated when different systems use different field structures or identifiers.

These improvements give data teams more flexibility when connecting, transforming, and enriching information across systems.

10. Better Currency Handling in Data Processing Engine

Data 360 is also adding more control over currency handling in the Data Processing Engine.

For organizations operating across countries, currencies, and business units, consistent currency processing is important for analytics and financial data.

Why it matters

Better currency handling can help businesses process financial information more consistently when data comes from multiple sources and markets.

The 5 Data 360 Updates I'd Highlight Most

Out of the 10 updates covered above, these five are particularly interesting for businesses looking at AI, analytics, data integration, and governance.

Final Thoughts

Salesforce Winter ’27 shows that Data 360 is moving beyond simply bringing data together. The latest updates focus on making connected data easier to analyze, govern, and use for AI and business processes.

Features such as advanced regression, unsupervised clustering, granular Data Graph governance, Snowflake Zero-Copy V2, and AI-assisted Document AI give businesses more flexibility to work with their data while maintaining control.

For organizations preparing for Agentforce, AI, analytics, or automation, the quality and structure of their data should be a priority. If your data is spread across multiple systems, contains duplicates, or needs to be migrated and prepared for Data 360, the right data strategy can make a significant difference.

Planning a Salesforce data migration or preparing your data for Data 360? Talk to a Salesforce expert to assess your current environment and build the right approach for your business.

Frequently Asked Questions

1. What is Salesforce Data 360 in Winter ’27?

Salesforce Data 360 is Salesforce’s data platform for connecting, unifying, analyzing, and governing business data. The Winter ’27 updates add new capabilities for predictive analytics, clustering, data governance, zero-copy connectivity, Document AI, and more.

2. What are the key Data 360 updates in Winter ’27?

Some of the most notable updates include advanced regression loss functions, unsupervised clustering, granular Data Graph governance, Snowflake Zero-Copy V2, Python transformations on zero-copy data, and the Data 360 MCP Server for Document AI.

3. How does unsupervised clustering work in Data 360?

Unsupervised clustering groups records based on patterns and similarities in the available data without requiring predefined labels. Businesses can use these groups for customer segmentation, pattern discovery, personalization, and predictive analytics.

4. What are Gamma and Quantile regression used for in Data 360?

Gamma regression can help predict positive, right-skewed outcomes such as revenue, customer spending, or claims. Quantile regression can predict a specific percentile, such as the 90th percentile of delivery times, rather than only estimating an average outcome.

5. What is Snowflake Zero-Copy V2 in Data 360?

Snowflake Zero-Copy V2 allows organizations to work with Snowflake data through Data 360 without unnecessarily duplicating the underlying data. This can help reduce data movement while allowing businesses to use external data in their broader data and AI workflows.

6. How does Data 360 improve data security in Winter ’27?

Granular data graph governance provides more detailed control over connected data. Organizations can use capabilities such as field pruning, data masking, and exclusion of unauthorised child data model objects to help protect sensitive information.

7. Can Data 360 run Python transformations on external data?

Yes. Code Extensions can run custom Python batch transformations on data backed by supported zero-copy connectors, allowing technical teams to apply custom processing without first copying the external data into Data 360.

8. What is the Data 360 MCP Server for Document AI?

The Data 360 MCP Server for Document AI allows compatible AI development tools to interact with Data 360 using natural language. Teams can use it to assist with tasks such as creating document schemas, configuring extraction, testing results, and publishing configurations.

9. When will the Salesforce Winter ’27 Data 360 features be available?

Data 360 follows its own release cadence, and Salesforce notes that Winter ’27 changes are generally included in the October 2026 release. As this release is in preview, organizations should consult the individual release notes to ensure that they are aware of the latest requirements and availability.

10. Why are these Data 360 updates important for AI?

AI depends on reliable, relevant, and governed business data. Winter ’27 strengthens several parts of that foundation, from connecting external data and discovering patterns to controlling access and processing unstructured information. This can help organizations build AI experiences on a more trusted data foundation.