⚡

Salesforce Certified Agentforce Specialist

Advanced

Validates the skills to design, build, and deploy autonomous AI agents using Agentforce — including topics, actions, prompt grounding, and governance.

72%

Passing score

105 min

Exam duration

Exam questions

35h

Study estimate

Topic Exams

Named exams are a Premium feature

Upgrade to Premium for structured Practice Exams and Topic Exams with retake tracking, plus unlimited practice questions — or buy just this certification.

Upgrade to Premium

Topics covered

AI Agents

35%

Agentforce Fundamentals

Core concepts: what an autonomous agent is and how Agentforce fits the platform.

AI Agents

Agent types, agent builder, and agent lifecycle.

12Q

Agent Actions

Standard and custom actions an agent can invoke.

10Q

Agent Topics

Scoping an agent's job using topics, instructions, and scope.

Einstein Copilot

Copilot actions and conversational assistant configuration.

Flow Integration

Invoking Flows as agent actions and passing data between them.

Apex Actions

Invocable Apex methods exposed to agents as custom actions.

Custom Copilot Actions

Building and registering custom actions for Copilot/Agentforce.

Best Practices

Design patterns for reliable, safe, and maintainable agents.

Prompt Engineering

20%

Prompt Builder

Building grounded prompt templates for agent responses.

11Q

Prompt Templates

Reusable prompt template types: flex, field generation, record summary.

10Q

Prompt Grounding

Merge fields, related lists, and Data Cloud grounding in prompts.

11Q

Retrieval-Augmented Generation (RAG)

Grounding generative responses in retrieved, trusted data sources.

11Q

Data 360 Fundamentals

20%

Data Cloud Integration

Grounding agent responses with harmonized Data Cloud data.

21Q

Model Builder

Bringing custom or external models into the Salesforce AI platform.

21Q

Testing, Deployment, & Maintenance

10%

Testing

Testing agent conversations, actions, and edge cases before go-live.

10Q

Deployment

Moving agent configuration across orgs and environments.

10Q

Governance & Observability

10%

Governance

Guardrails, approval processes, and responsible deployment policy.

AI Security

Permission sets, shield, and secure configuration for agent access.

Einstein Trust Layer

Data masking, toxicity scoring, and secure data handling for agents.

Agent Analytics

Monitoring agent utterances, conversation success, and deflection.

Multi-Agent Orchestration

Coordinating multiple specialized agents, agent-to-agent handoff, and interoperability protocols.

11Q