See What’s New: Data 360 and Tableau
Learning Objectives
After completing this unit, you’ll be able to:
- Describe the latest features and enhancements for Data 360 and Tableau.
Learn about the new features in Data 360 and Tableau that make it easier than ever to integrate AI into everyday workflows. The right tools and context can make or break how AI functions in the real world. Salesforce gives you access to accurate data, at the right time, in the right context to confidently automate tedious tasks and focus on real value.
Data 360
Agentforce Coworker
Most AI initiatives don’t fail because the model is bad. They fail because the model lacks trusted enterprise context: your workflows, your compliance rules, the customer insights that run your business. Agentforce Coworker is your autonomous AI teammate that knows your business from day one. Agentforce Coworker is already connected to opportunities, cases, accounts, pipeline, service history, and your entire trusted enterprise context, with the permissions, governance, and controls you expect from Salesforce. It takes action on your behalf, orchestrating agents, executing complex tasks, and driving work forward from wherever your team works.
Data 360 Headless
AI can answer questions, but without business context it doesn’t understand your customers, your business rules, or how your company operates. Now, Data 360 Headless makes the context that’s locked inside Salesforce available anywhere. The Data 360 MCP Server securely connects that context to AI applications like ChatGPT, Slackbot, and any MCP-compatible agent, without custom integrations or moving your data. Prebuilt Skills package your business logic so every agent can reason and take action exactly as you’d expect.
Extend trusted business context to any agent, app, or surface beyond Salesforce, with governance and permissions enforced at the source. Your teams move faster, your agents perform better, and your data stays trusted everywhere you work.
Notebook AI Deep Research
In the past, to get answers to complex business questions, users had to dig through CRM records, documents, enterprise systems, and the web–a fragmented process that was easy to get wrong. Notebook AI’s Deep Research changes that. Users now can ask a question, review, and approve a plan, and the system runs through the key sources on its own, including Data 360 indices, so nothing important gets missed. The result is a comprehensive, cited report grounded in your actual business data, structured, and ready to act on. Every report is saved in Notebook AI, turning one-off research into a resource the whole team can query and build on.
RLS (Role-Level Security) for Data Graph
AI agents built on Data Graph now enforce role-level security, alongside existing object and field-level controls. Restrictions set on underlying data follow all the way through to the agent, so every agent only works with the records it’s permitted to see. Admins define deny policies directly on a Data Graph using field-based conditions that can filter specific records within a node or block access to the entire graph when a condition isn't met. Policies are enforced before any data reaches the agent, independent of anything set on the source objects.
Zero Copy Data Sharing with Google BigQuery
Zero Copy Data Sharing with Google BigQuery gives every Data 360 customer near real-time access between Data 360 and BigQuery—no pipelines, no copies, no regional restrictions. Building on our Zero Copy foundation, the connection now runs through Google Analytics Hub to extend availability to every Data 360 tenant in any AWS region.
Tableau
Tableau Next and Tableau MCPs Available in Slackbot
Most analytics workflows force users to either wait for a scheduled report or stop everything and dig through a dashboard. Tableau Next and Tableau MCP in Slackbot help users get results instantly with two complementary capabilities—Proactive Briefings that surface the right insights before you ask, and Conversational Q&A that lets you investigate the moment a question surfaces.
Admins configure the connection once; analysts curate the semantic models and business context that power governed answers. The result is an analytics experience that meets business users where they already work: users can move from insight to investigation to action in a single Slack thread, without switching tools or waiting on analyst support.
Tableau Agent: Conversational Analytics Enhancements
When users ask complex analytical questions such as period-over-period comparisons, trend analysis, and top driver breakdowns, even well-built data environments return inconsistent or incomplete answers. Tableau Agent Conversational Analytics Enhancements address this with a new query generation engine that delivers highly accurate, consistent results across a broader range of question types, including advanced visualizations like scatter plots, doughnut charts, and heat maps.
A “chain of thought” UI surfaces step-by-step reasoning as the agent works through complex problems. And upgraded multi-turn memory keeps context alive across an entire conversation. These enhancements give every user the accurate, transparent analytical foundation that autonomous action requires.
Semantic Model Builder: Smart Canvas
The current semantic model builder cannot gracefully handle enterprise scale. Once object counts climb past 50, the builder breaks down, and even experienced authors avoid making changes for fear of breaking production. Smart Canvas replaces the rigid wizard-style experience with a modern, fluid design surface.
Auto-layout algorithms organize objects into domain-based clusters, inline metadata editing removes the extra clicks that pull authors out of flow, and a “Try Now” banner makes adoption instant with no configuration required. Data modelers can now navigate, edit, and iterate on enterprise-scale models with confidence. Agents and analytics are only as good as the models beneath them, and this builder makes it easy to scale models for success.
Hosted Tableau MCP
Connecting AI agents to enterprise analytics has required standing up and maintaining your own MCP server, a burden most teams don’t want and can’t scale. Hosted Tableau MCP now lets you run a fully managed MCP server natively in Tableau Cloud, so you can expose workbooks, data sources, and metrics as structured tools that any MCP-compatible agent can call.
Admins no longer need to provision, secure, or operate separate infrastructure to bring Tableau context into an AI workflow. With OAuth 2.1 authentication, built-in governance, and a discoverable endpoint at mcp.tableau.com, any agent—including Agentforce, Claude, or ChatGPT—connects to trusted Tableau analytics with the reliability and controls an enterprise requires. Ground AI in governed, real-time data without adding operational overhead.
CRM Analytics MCP Server (Beta)
Developers and analysts using Claude, Cursor, or GitHub Copilot have had no direct path to CRM Analytics data; every AI-assisted workflow required manually extracting data first, breaking context, and introducing risk. The CRM Analytics MCP Server (Beta) connects those tools directly to your CRMA org. Just point your AI tool at the MCP server endpoint and it works, no custom connectors, no integration code, no new permissions to configure. Agents can run SAQL queries, list dashboards, and explore datasets in real time, all under the user’s existing access controls and with no data leaving the platform.