As organizations race toward the AI and agentic era, enthusiasm often outpaces architectural reality. The true bottleneck to successful AI adoption is rarely the algorithm, it is the state of the underlying data estate. In this session, we explore why rigorous data governance is the ultimate prerequisite for sustainable AI transformation. Featuring a practitioner and customer-centric perspective, we examine how establishing clear data definitions, ownership, relevance, and importance; data integrity; and robust architectural alignment between systems of record and systems of engagement paves the way for trusted, actionable insights and autonomous business scale.

Data lake, warehouse, lakehouse, Bronze/Silver/Gold, golden records, data intelligence - the modern data landscape is full of terms that are often misunderstood, or used interchangeably. This session demystifies these concepts, explains how they fit together, and explores their role in building trusted, well-governed enterprise data environments.

Salesforce has what is arguably the enterprise’s most critical asset: its customer data. As enterprise AI transforms the platform from a traditional CRM into an active AI engine, Salesforce architects face unprecedented challenges unifying massive amounts of data across platforms without compromising security or governance. This keynote explores the architectural foundation needed for trusted, scalable, AI-ready data to better serve data teams.

You’ll learn about:

• Business & security context for AI
  
• AI data cost optimization
• Scalable data pipelines
• Data freshness

• Large data volume (LDV) management
• Zero-copy integration
• Agent observability

This session shares a practical, four-phase roadmap for Salesforce architects:

• Make APIs agent-ready
• Adopt MCP for plug-and-play interoperability
• Move from scripted workflows to governed orchestration with humans in the loop
• Build the observability to treat agents as an auditable digital workforce

Along the way, we show how a semantic data layer and governed context let agents scale beyond the pilot.

Agentforce agents read Salesforce data through a user's permissions, either a dedicated agent user or the person they are helping, so an agent can reach whatever that user can reach. Salesforce enterprise architects will trace one customer-facing agent to find where its data boundary actually sits and why verifying a customer does not narrow it. Architects will leave with three checks covering agent user permissions, the sharing model, and the context each action runs in, which they can use to test their own agents.

Discover architecture patterns for governing multi-agent AI systems. Learn approaches for orchestration, trust, escalation, observability, and secure Agent2Agent collaboration at enterprise scale.

Join us for an exploration of the evolving landscape of AI data security. As organizations integrate more AI solutions, the risks associated with data consolidation, visibility, and access control are increasing. This session will provide actionable insights into the top five security best practices for AI, focusing on:

• Minimizing Attack Surfaces: Implementing zero-copy architectures to reduce data duplication
• Granular Data Control: Strategies for selecting exactly what information is exposed to AI models
• AI-Driven Masking: Techniques for maintaining confidentiality and enforcing encryption when sharing data with AI
• Centralized Visibility: Building architectures that unify management without sacrificing oversight
• Business Continuity: Ensuring BCDR readiness to protect critical AI-driven business processes

Grounding gives AI agents facts. Context gives them meaning - and that's where the advantage lies. Every enterprise has centers of gravity: bodies of data that pull applications, decisions and priorities towards them. Agents now orbit them, too.

Drawing on 15 years in Salesforce and customer-led transformation across multiple industries, this session connects data gravity and the agentic enterprise through two composite stories.

Attendees will learn how to:

• Identify their own centers of gravity
• Measure data by its consequence rather than its volume
• Decide where data should be mastered, moved, or copied
• Match deterministic and probabilistic reasoning to the weight of each decision

This includes practical principles for data and AI decisions, a test for executives and architects on each, and actions you can take back to your business and apply the very next day.

Building a Data 360 solution no longer requires extensive manual effort. In this session, you will see how AI tools using Model Context Protocol (MCP) integrations with Salesforce can drive 70-80% of the build process, with only occasional human oversight. We will walk through packaging Data 360 components, navigating common metadata dependencies, and wiring everything into a repeatable deployment pipeline. You will leave with practical strategies to go from AI-assisted development all the way to successful, automated delivery across QA and production environments.

Discover best practices for implementing fine-grained data governance, zero-trust access controls, and end-to-end pipeline observability across complex Salesforce data flows to ensure compliance and prevent data drift.

This session shows how to turn Salesforce data into working, governed applications in a fraction of the usual time. Attendees will see how teams can go from a business question to a secure, scalable app without long development cycles, while staying in control of their data. You’ll learn about:

• Rapid App Design Principles: how to structure data, logic and UI to turn CRM data into usable apps quickly
• Built-in Governance & Security: keeping access controls, compliance and data protection in place as apps multiply
• Scaling & Integration Best Practices: keeping apps fast and reliable, and connecting them with your wider data and AI ecosystem