Salesforce Agentforce Training for CRM Teams

Agentforce is Salesforce's agent platform, built into the CRM your teams already use. Its agents are made of topics and actions. Grounding can draw from accounts and cases as well as opportunities. CRM and sales operations teams learn beside service staff to scope an agent around a defined use case. They ground it in the right data and add the guardrails and human checkpoints that keep a live agent trustworthy inside your org. The syllabus stays Salesforce-specific throughout and finishes with a governed, measured rollout plan.

modules
6
hours
13
Contact us

This training spends two days on Agentforce specifically. Your team builds and scopes an agent in Agent Builder before grounding it in your own Salesforce data. It then designs service and sales workflows with clear handoff points to a human. Guardrails and tests keep it inside its intended scope through change control. A closing module carries a working pilot through to a governed, measured rollout.

Agentforce is Salesforce's agent platform. Its agents use topics and actions in Agent Builder while drawing on accounts and cases as well as opportunities. They run inside Service Cloud and Sales Cloud conversations. That gives the work different constraints and data risks from general agent literacy. Turning it on is quick. Configuring it to behave well across a live book of business takes care. A team must scope when an agent engages and ground its actions in the right objects and permissions. It also decides what may write back before a human checks it. Those choices determine whether the agent helps or quietly creates new problems in the CRM.

For agent concepts that apply beyond Salesforce, see our AI Agents training. It covers reasoning and tool use alongside governance across platforms. This page stays Agentforce-specific. Every module here is adapted to your Salesforce edition and objects, including existing automation, before the first session. The AI training catalog holds the platform-neutral programs.

Agentforce puts a live agent inside your CRM

An agent that reads and writes CRM data touches live accounts, cases, and opportunities, the same records your team works from every day. Get the configuration wrong and the mistake is live immediately. Configuring it well matters more here than in a low-stakes pilot elsewhere.

The platform is easy to turn on, hard to configure well

Standing up an agent in Agent Builder takes an afternoon. Scoping its topics and actions takes deliberate practice most teams never get, especially when grounding has to hold across real conversations.

What happens when guardrails or grounding are wrong?

An Agentforce agent is only as trustworthy as the data it is grounded on and the guardrails around its actions. Get either wrong and you get a confident agent giving the wrong answer or taking the wrong action.

Service and sales teams need different agent designs

A case-triage agent and an opportunity-research agent solve different problems with different data and different risk. Treating "add an agent" as one decision produces a design that fits neither team well.

An ungoverned pilot never scales

Early enthusiasm for a working agent can outrun the approval process behind it, and deciding who owns changes to topics and actions, plus how a rollout gets measured, has to happen before broad rollout begins.

  1. Where Agentforce fits: agentic AI inside Salesforce

    90 minBeginner

    What agentic AI means inside a Salesforce org, and where Agentforce sits next to the Flows, Apex, and bots your team may already run. One working picture of what Agentforce can and can't do, before anyone builds an agent against live CRM data.

    • What topics and actions mean for an agent inside Agentforce
    • How Agentforce differs from Flow and Apex or bot automation
    • Where Agentforce fits inside Service Cloud and Sales Cloud conversations
    • What it does well, and where a human rep or agent still has to decide
    • A single low-risk first experiment to build intuition
  2. Agents built from topics and actions in Agent Builder

    120 minBeginner

    This module works directly inside Agent Builder. Participants create an agent, define its topics, wire up the actions it can take, and test the result in a sandbox before it ever reaches a live conversation.

    • Creating an agent and its topics in Agent Builder
    • Standard and flow-based actions alongside Apex-invoked work
    • Scoping a topic so the agent knows when to engage, and when not to
    • Writing instructions that keep an agent inside its intended job
    • Testing an agent in a sandbox before it reaches a live conversation
    • Where this training ends and Salesforce implementation or custom development begins
  3. Grounding Agentforce in your CRM data

    150 minIntermediate

    An agent's answers and actions are only as good as their grounding, so this module uses the right Salesforce objects and records inside the permissions your org already enforces.

    • Grounding actions and answers in the right Salesforce objects and records
    • Respecting field- and record-level permissions so an agent sees only what its user should
    • Choosing approved data sources for grounding, and keeping out what is not approved
    • Avoiding ungoverned grounding sources
    • Testing grounded responses against approved records in a sandbox before go-live
  4. Designing service and sales workflows with Agentforce

    150 minIntermediate

    From a business brief to a working agent. Participants design a Service Cloud or Sales Cloud workflow end to end, with clear handoff points to a human before anything writes back to a record.

    • Service Cloud patterns for case triage and grounded replies with escalation
    • Sales Cloud patterns for account research and meeting prep behind CRM updates
    • From a business brief to a scoped topic and action list
    • Human checkpoints for any action that writes back to a record
    • Building one governed workflow on your own Salesforce environment
  5. Trust controls built from guardrails and testing

    150 minIntermediate

    Speed without guardrails is a liability inside a CRM. Teams need a concrete discipline for agent instructions and tests, followed by monitoring. That practice decides whether the agent earns their trust.

    • Instructions and guardrails that keep an agent inside its intended scope
    • Accuracy and tone rubrics, with correct escalation tested before go-live
    • Salesforce's built-in trust and data-handling controls, and what they do not replace
    • Human review and escalation paths for a conversation the agent should not finish
    • Change control across topics and actions, including their instructions
    • Catching drift after launch
  6. Governed rollout and measuring adoption

    120 minAdvanced

    One working pilot becomes an agent your CRM team trusts and measures. The rollout path defines who owns the agent and keeps it governed. Honest metrics show adoption rather than activity.

    • Choosing a pilot use case and success measures before a broader rollout
    • A rollout path: sandbox pilot, limited team, then phased rollout
    • An ownership model spanning admins and CRM ops beside business stakeholders
    • Adoption and quality metrics
    • Revisiting scope over time
    • A rollout plan you leave with, built on your own Salesforce environment

What you will learn

  • Scope an Agentforce agent that matches a current use case
  • Ground agent responses in the right Salesforce data
  • Design a service or sales workflow with clear human checkpoints for actions that write back to the CRM
  • Reliable guardrails backed by tests
  • Decide who approves changes
  • Adoption metrics that hold up
  • Know when a use case needs an agent
  • Keep customer data inside policy as agents act

Who should attend

  • Salesforce administrators and CRM platform owners
  • Sales operations and RevOps teams
  • Service Cloud and Sales Cloud managers
  • CRM and product owners scoping their own Agentforce rollout
  • IT and Salesforce Center of Excellence stakeholders
  • Team leads piloting Agentforce for their department

Agentforce labs take two days, usually split into half-day sessions so delivery schedules are not disrupted. Onsite and live-online formats cover the same syllabus, and the lab work runs in a Salesforce sandbox or your own org where you allow it.

Format
Onsite or live online
Duration
2 days (about 13 hours, can be split into half-day sessions)
Group size
Up to 16 participants per group
Materials
Agent scoping template, guardrail checklist, and rollout plan
Language
English or Turkish
Certificate
Certificate of completion

Zeo started in 2011 and now works out of San Francisco, Istanbul, Ankara, and Lisbon. We run Copilot Academy and organize Digitalzone, an international digital marketing conference. This program draws on the 10+ years of consulting and training work behind that, applied to corporate AI adoption.

  • 2011founded in Istanbul
  • 10+years of consulting and training experience
  • 3offices: San Francisco, Istanbul, Ankara, Lisbon
Send over your Salesforce org and service or sales use cases, including where Agentforce already runs today. We will use those details to design the labs and rollout plan for your CRM.
Contact us
Illustrated figure working on a laptop surrounded by floating tool windows