The Salesforce Certified Agentforce Specialist exam has been completely redesigned to reflect the platform’s shift toward intelligent, data-driven, and interoperable agents. The new version focuses on how agents think, interact, and collaborate within a connected enterprise.
The previous guide centered on configuration and product-specific features. The new guide highlights the broader architecture of Agentforce and its alignment with Salesforce’s enterprise AI vision.
1. A Completely New Domain: AI Agents (35%)
The most significant addition is the introduction of the AI Agents domain. It replaces the older sections that covered Agentforce Concepts, Service Cloud, and Sales Cloud. This domain consolidates everything that defines how agents function, communicate, and operate securely.
It introduces new objectives such as managing deterministic behavior using filters and variables, explaining how the reasoning engine powers Agentforce, and selecting the correct topics and actions based on agent type. It also introduces the Agent User concept to manage security and permissions for Employee, Service, and Sales agents. Another addition is configuring agents across channels such as Slack, email, and digital experience.
This new section represents Salesforce’s move from isolated agent use cases to a cohesive multi-channel framework where every agent type has a defined operational role.
2. Prompt Engineering Refined and Reduced (30% → 20%)
The Prompt Engineering domain still plays a major role but now also focuses on prompt quality. Its weighting has decreased from 30 to 20 percent.
Two new topics appear: considerations for creating prompt templates using field generation and flex types, and best practices for writing effective prompts. These changes emphasize prompt reliability and consistency across use cases. Prompt Builder remains central, but the focus has shifted from creating prompts to engineering them with measurable intent and context.
3. Data Cloud for Agentforce Deepens Its Scope (20%)
Although it retains the same weighting, this domain has been completely rewritten. The new objectives require understanding of Agentforce Data Library types, unstructured data processing through chunking and indexing, and using retrievers such as individual and ensemble.
Candidates must now understand search type selection (keyword, vector, or hybrid) and how it affects response accuracy. This reflects Salesforce’s focus on grounding agents in enterprise data, particularly unstructured sources, to improve reliability and relevance.
4. Development Lifecycle as a Dedicated Domain (20%)
Previously included in the broader Agentforce Concepts section, lifecycle management is now a standalone domain. It validates a professional’s ability to build, test, deploy, and monitor agents across environments.
It introduces structured processes for testing agents in the Agentforce Testing Center, deploying safely from sandbox to production, and tracking adoption to ensure that deployed agents deliver value. This acknowledges that Agentforce development now mirrors the rigor of application lifecycle management.
5. Multi-Agent Interoperability Added (5%)
The new Multi-Agent Interoperability domain is the clearest signal of where Salesforce is heading. It introduces three architectural components that define how agents collaborate and extend functionality:
- Model Context Protocol (MCP) connects agents to external systems and data sources.
- Agent-to-Agent Protocol (A2A) enables agents to exchange context, delegate tasks, and coordinate responses.
- Agent API allows programmatic access so external systems or websites can communicate with AI agents directly from a REST API.
This section represents Salesforce’s vision of connected, intelligent systems that exchange data and reasoning securely across boundaries.
6. Removal of Product-Specific Segments
Older sections dedicated to Service Cloud and Sales Cloud have been retired. Their relevant material is now integrated into the AI Agents domain. The new exam guide recognizes that Service and Sales agents are part of a unified ecosystem rather than isolated product-based implementations.
Final Insight
The updated Agentforce Specialist certification moves beyond configuration and product familiarity. It now validates expertise in agent reasoning, governance, and interoperability.
Professionals preparing for this exam must demonstrate their ability to build agents that think contextually, access data responsibly, and work together as part of a larger AI system. Salesforce has reframed Agentforce not as a feature but as an architectural foundation for enterprise intelligence.
Ready to go deeper?
Get the full breakdown, study plan, and resources in our Salesforce Agentforce Specialist Certification Guide. Then prep with the latest Study Guide and Practice Exams—updated for the new domains and weightings—so you walk into exam day confident.
