Change Log
View the Framework to explore the Well-Architected Framework for Salesforce.
This page tracks all updates to Architecture Center content. Check back regularly to stay current with the latest guidance.
Framework
| Published | Well-Architected Framework: The principled foundation for architectural quality on Salesforce, organized around five pillars (Trust, Reliability, Operational Excellence, Resource and Cost Optimization, and Fairness) that apply across traditional and agentic solutions. |
| Published | Trust: Design secure, compliant solutions through the Shared Responsibility Model. Covers identity and access, data protection, compliance, secure development, and incident response. |
| Published | Reliability: Design available, resilient solutions that scale within governor limits and recover from disruption. Covers scalable data models, fault-tolerant transactions, availability monitoring, and disaster recovery. |
| Published | Operational Excellence: Build observable solutions managed through defined procedures, automation, and continuous improvement. Covers observability, deployment safety, DevOps, and incident readiness. |
| Published | Resource and Cost Optimization: Design solutions that return maximum value per cost through efficient use of platform capacity and deliberate investment in what you buy. Covers TCO analysis, right-sizing, governor limits, and licensing discipline. |
| Published | Fairness: Ensure solutions are accessible, non-discriminatory, and transparent with human oversight of automated decisions. Covers WCAG 2.2 AA accessibility, bias detection, transparent decisions, and ethical governance. |
| Published | The Agentic Enterprise: Companion to the Well-Architected Framework explaining what changes when solutions become agentic. Covers how humans, agents, and platforms work together inside governed business systems, and points to pillar-by-pillar guidance. |
| Published | Trust for Agentic Enterprise: Govern what agents may do and audit what they have done. Covers agent identity, scoped permissions, and comprehensive audit trails for autonomous action. |
| Published | Reliability for Agentic Enterprise: Design reliable Agentforce solutions that anticipate non-deterministic failures. Covers multi-tier fallback chains, graceful degradation, and automatic recovery from disruption. |
| Published | Operational Excellence for Agentic Enterprise: Observe probabilistic agentic systems and manage them through the Agent Development Lifecycle. Covers agent observability, evaluation across quality dimensions, and lifecycle governance. |
| Published | Resource and Cost Optimization for Agentic Enterprise: Keep agent consumption efficient and spend aligned with value. Covers bulkified Actions, disciplined context windows, consumption-based pricing models, and TCO modeling for variable agent workloads. |
| Published | Fairness for Agentic Systems: Ensure agent-driven experiences are explainable, contestable, and equitable. Covers bias detection for consequential agent decisions, substantive human oversight, and the Einstein Trust Layer. |
| Published | Change Log: Tracks all updates to Architecture Center content, showing what's new, what's changed, and what's been removed. |
Fundamentals
| Enhanced | Asynchronous Processing on the Lightning Platform: Updated Well-Architected Framework references from the legacy Trusted/Easy/Adaptable model to the current five-pillar taxonomy. |
| Enhanced | Platform Sharing Architecture: Updated Well-Architected Framework references from the legacy Trusted/Easy/Adaptable model to the current five-pillar taxonomy. |
| Enhanced | Salesforce Data Synchronization: Updated Well-Architected Framework references from the legacy Trusted/Easy/Adaptable model to the current five-pillar taxonomy. |
| Enhanced | Salesforce Data Virtualization: Updated Well-Architected Framework references from the legacy Trusted/Easy/Adaptable model to the current five-pillar taxonomy. |
| Enhanced | Data 360 Document AI: Updated Well-Architected Framework references from the legacy Trusted/Easy/Adaptable model to the current five-pillar taxonomy. |
Decision Guides
| Enhanced | Building Forms: Updated Well-Architected Framework references from the legacy Trusted/Easy/Adaptable model to the current five-pillar taxonomy. |
| Enhanced | Data Integration: Updated Well-Architected Framework references from the legacy Trusted/Easy/Adaptable model to the current five-pillar taxonomy. |
| Enhanced | Event-Driven Architecture: Updated Well-Architected Framework references from the legacy Trusted/Easy/Adaptable model to the current five-pillar taxonomy. |
Fundamentals
| Published | Data 360 Document AI: Extracting structured data from unstructured documents and grounding it in Data 360, making document content available across analytics, automation, and Agentforce. |
| Published | Architectural Patterns for Field Service: Optimizing Contractor Sharing and Visibility at Scale: Compares territory-based and account-based sharing patterns for governing external contractor networks in Field Service. Covers how to balance platform security with scheduling performance and build the data foundation for AI-driven scheduling with Agentforce. |
Decision Guides
| Enhanced | Data 360 Interoperability: Updated to new decision guide format; split Zero Copy federation into three distinct patterns (Live Query, Accelerated Query caching, and File Federation), with guidance on when to use each based on query frequency, latency, and cost. |
Fundamentals
| Published | Agentic Integration Patterns: Architectural blueprint for the integration patterns agents depend on. Covers how to design integrations that handle ambiguity, support safe repetition, and fail recoverably, and how the Salesforce platform delivers each pattern. |
Reference Architectures
| Enhanced | Reference Diagrams: Complete refresh of all diagrams to improve accessibility and readability. Updated product names. Other enhancements. |
Fundamentals
| Enhanced | Integration Patterns: Legacy tools deprecated in favor of Flow and Platform Events. GraphQL and Pub/Sub API replace SOAP. OAuth 2.0 is now standard. Data Cloud expanded to include Data 360 and Zero Copy patterns. Event Relay bridges Salesforce with AWS EventBridge. |
| Published | Salesforce Data Synchronization: Comprehensive guide to data synchronization patterns on Salesforce. Covers batch sync, real-time sync, bidirectional sync, and conflict resolution strategies for keeping data consistent across systems. |
| Published | Salesforce Data Virtualization: Guide to accessing external data in real-time without copying it into Salesforce. Covers Salesforce Connect, External Objects, and zero-copy integration patterns with Data Cloud. |
Fundamentals
| Published | Hybrid Reasoning with New Agentforce Builder and Agent Script: Guide to combining LLM reasoning with deterministic logic in Agentforce. Shows when to use Builder's visual tools versus Agent Script for complex workflows requiring both AI flexibility and business rule precision. |
Decision Guides
| Published | Determining Agentic and Traditional Workflow Automation: Decision guide for choosing between autonomous agents and traditional automation. Compares Agentforce, Flow, Apex, and other automation tools based on use case characteristics like variability, human oversight, and integration complexity. |
Fundamentals
| Published | Leveraging MuleSoft and Informatica: Better Together: MuleSoft for application connectivity and APIs; Informatica for data integration, quality, and MDM. How to pair them for trusted scale and agentic workflows. |
Decision Guides
| Enhanced | Data 360 Provisioning Decision Guide: Updated to new decision guide format. |
Decision Guides
| Enhanced | Data 360 Provisioning: Updated to new decision guide format. |
Decision Guides
| Enhanced | Build Forms: Updated to new decision guide format. |
| Enhanced | Building Forms Decision Guide: Updated to new decision guide format; updates for product/feature currency. |
Fundamentals
| Published | End-User Identity Propagation in Agents: How to maintain end-user identity context when agents make API calls to external systems. Covers OAuth token exchange, named credentials, and patterns for ensuring agents act with appropriate user permissions rather than service account permissions. |
Decision Guides
| Enhanced | Async Processing: Updated to new decision guide format. |
Fundamentals
| Published | Data 360 Clean Rooms: Guide to privacy-preserving data collaboration using Data Cloud Clean Rooms. Enables secure data sharing and analytics across organizational boundaries without exposing raw customer data. |
Decision Guides
| Enhanced | Event-Driven Architecture: Updated to new decision guide format. |
| Enhanced | Data 360 Interoperability: Updated to new decision guide format. |
| Enhanced | Data 360 Interoperability Decision Guide: Updated to new decision guide format. |
Reference Architectures
| Published | Zero-Copy Data Integration Architecture Between Salesforce Data 360 and Databricks: Reference architecture for connecting Data Cloud to Databricks lakehouse without data movement. Leverages zero-copy federation for real-time analytics and ML model training on unified customer data. |
| Published | Data 360 Batch Activation and Event Driven Action Architecture: Reference architecture showing scheduled batch activations and real-time event-driven actions from Data Cloud. Demonstrates how to orchestrate both batch segments and streaming activations to downstream systems. |
Fundamentals
| Published | Asynchronous Processing on the Lightning Platform: A guide comparing the products, tools, and approaches for asynchronous processing on the Salesforce Platform. Explains how each technology works under the hood, where async fits and where it doesn't, and the anti-patterns to avoid—complementing the Asynchronous Processing Decision Guide. |
Fundamentals
| Published | Content: High Availability and Disaster Recovery : MuleSoft: Guide to architecting resilient MuleSoft integrations. Covers runtime redundancy, CloudHub availability zones, disaster recovery strategies, and failover patterns for mission-critical integration flows. |
Decision Guides
| Enhanced | Data Integration: Updated to new decision guide format. |
| Enhanced | Record Triggered Automation: Updated to new decision guide format. |
Decision Guides
| Enhanced | Event Driven Architecture Decision Guide: Updated to new decision guide format. |
Decision Guides
| Enhanced | Data Integration: Added Data 360 and Data Cloud One as options for data integrations. |
Reference Architectures
| Published | Customer Zero Data platform Reference Architecture: Reference architecture showing how Salesforce uses Data Cloud internally (Customer Zero). Demonstrates real-world implementation patterns for enterprise-scale data unification, identity resolution, and analytics. |
| Published | Zero-Copy Data Integration Architecture Between Salesforce Data 360 and AWS: Reference architecture for federating Data Cloud with AWS data lakes (S3, Redshift) without copying data. Enables unified queries across Salesforce and AWS using zero-copy connectors and Data Cloud query federation. |
| Published | Zero-Copy Data Integration Architecture Between Salesforce Data 360 and Snowflake: Reference architecture for connecting Data Cloud to Snowflake Data Cloud without data duplication. Leverages zero-copy federation for unified customer 360 views across both platforms. |
| Published | Zero-Copy Data Integration Architecture Between Salesforce Data 360 and GCP: Reference architecture for federating Data Cloud with Google Cloud Platform (BigQuery, Cloud Storage). Demonstrates zero-copy integration patterns for analytics and ML workloads spanning Salesforce and GCP. |
Fundamentals
| Published | Data 360 Security Architecture: Comprehensive guide to securing Data Cloud implementations. Covers data classification, encryption at rest and in transit, access controls, audit logging, and compliance considerations for unified customer data platforms. |
| Published | Data 360 Integration Patterns: Pattern catalog for integrating Data Cloud with Salesforce clouds and external systems. Covers ingestion patterns, activation patterns, federation patterns, and best practices for building a unified customer data architecture. |
Fundamentals
| Published | Mulesoft Agent Fabric - Deep Dive: Deep dive into MuleSoft Agent Fabric architecture. Introduces the agent network concept and presents patterns for building intelligent, event-driven integration flows that combine traditional APIs with autonomous agent capabilities. |
| Enhanced | Architecting the Agentic Enterprise with Mulesoft: Added Integration Intelligence section. |
Fundamentals
| Published | The Agent Development Lifecycle: A methodology for building autonomous agents that adapts the traditional SDLC to their non-deterministic nature. Covers the five ADLC phases—Ideation and Design, Development, Testing and Validation, Deployment, and Monitoring and Tuning—with pro-code implementation guidance on Agentforce. |
| Published | The Agentic Enterprise - The IT Architecture for the AI-Powered Future: A point of view on the IT architecture enterprises need over the next 3–5 years to capture the value of an agentic workforce. Outlines how the traditional five-layer IT architecture must evolve to support autonomous agents, with a reference architecture for CIOs, CDOs, and IT leaders. |
| Published | Agentic Patterns and Implementation with Agentforce: A taxonomy of the primary AI agent types—Conversational, Proactive, Ambient, Autonomous, and Collaborative—with CRM use cases and an architectural blueprint for building each on Agentforce using Flow, Apex, Data 360, Agent2Agent (A2A), and Model Context Protocol (MCP). |
| Published | Enterprise Agentic Architecture and Design Patterns: A pattern-based methodology for designing multi-agent solutions, inspired by object-oriented design patterns. Presents agentic patterns from simple intent detection to multi-agent separation of concerns and UX patterns, with diagrams, usage guidance, and examples of composing them into holistic agentic experiences. |
| Published | Data 360 Architecture: A guide to the architecture, capabilities, and design principles of Data 360, Salesforce's cloud-native, metadata-driven data platform. Covers how it unifies siloed enterprise data for analytics, AI, and agentic applications, with deep dives into interoperability, security, governance, real-time activation, and Clean Rooms. |
| Published | Architecture Basics: An introduction to architecting on the Salesforce Platform's multitenant, metadata-driven architecture. Covers the three essentials every architect must understand: transactions and governor limits, the metadata-versus-data distinction, and platform APIs. |
| Published | Platform Multitenant Architecture: An under-the-hood technical brief on the multitenant, metadata-driven software architecture at the heart of the Salesforce Platform. Explains how the platform delivers reliable, scalable, and easily customizable applications across hundreds of thousands of tenants. |
| Published | Platform Sharing Architecture: A guide to architecting the Salesforce sharing model to meet current and future data access requirements. Covers record-level access components, sharing model use cases, real customer sharing solutions, and troubleshooting guidance for advanced admins and architects. |
| Enhanced | The Salesforce Platform - Transformed for Tomorrow: An under-the-hood look at the ground-up transformation of the Salesforce Platform for the AI era. Explores the architectural enhancements—across infrastructure, data, real-time scale, security, and resilience—that underpin Agentforce and keep the platform ready for next-generation agentic applications. |
Fundamentals
| Published | Step-Based Async Framework: Framework for running modular, scalable asynchronous jobs on Salesforce. Provides patterns for breaking long-running processes into steps with built-in observability, error handling, and resume capability for evolving enterprise needs. |
Fundamentals
| Published | Architecting the Agentic Enterprise with Mulesoft: Point of view for architecting the intelligent enterprise using MuleSoft. Explores how to combine traditional integration patterns with autonomous agents and AI-powered workflows for next-generation business automation. |