Salesforce Service Cloud: The Agentforce Service Guide

What Is Salesforce Service Cloud (Now Agentforce Service)?

Salesforce Service Cloud is the post-sale customer service product in the Salesforce ecosystem. It exists alongside Sales Cloud (pre-sale), Marketing Cloud (outreach), and the industry clouds (FSC, Health Cloud, etc.). Where Sales Cloud is built around the Lead and Opportunity, Service Cloud is built around the Case — a record of one customer’s question, complaint, or issue.

As of the Spring ’26 release, Salesforce officially rebranded Service Cloud as Agentforce Service. The current help documentation describes cases as “the backbone of Agentforce Service (formerly Service Cloud).” The product itself, the objects, the APIs, and the licensing did not change — what changed is how Salesforce wants you to think about it: not as a CRM cloud with bolted-on automation, but as a platform where human service reps and AI service agents share the same case queue.

The four pillars of Service Cloud

Strip away the marketing and Service Cloud reduces to four functional pillars. Everything else — Voice, Messaging, Field Service, Agentforce — is built on top of these:

Salesforce sells Service Cloud across five editions — Starter Suite, Pro Suite, Enterprise, Unlimited, and Einstein 1 Service — with each higher tier unlocking more channels, more automation, and more AI. Feature availability by edition is documented in the official Supported Editions reference, which should be the source of truth for any “does my org have this?” question.

The Case Object: The Heart of Salesforce Service Cloud

The Case object is a standard Salesforce object that ships with every org — but it only becomes the center of a real support practice when Service Cloud features are turned on. Out of the box, Cases support all the things you’d expect: a Subject, Description, Status, Priority, Origin, owner, related Account, related Contact, and a feed of activities.

Where Service Cloud extends the Case is in everything around it:

Case-related objects every admin should know

Two other features are worth calling out because they catch admins off guard:

Case Hierarchies. A case can have a parent case, which is how organizations model “this complaint is the same as the one I filed last week” or “this is a sub-task of a larger investigation.” Hierarchies are independent of Incidents and don’t share severity rollups.

Case Merge. Lightning Experience supports merging up to three duplicate cases into one master record, similar to Lead and Account merge. The merge consolidates Case Comments, Emails, and most related lists onto the surviving record.

Capturing Cases from Every Channel

A customer issue only becomes a Case when it lands in Salesforce. Service Cloud supports a long list of intake channels, each with its own setup, license requirements, and routing behavior. The main ones, from oldest to newest:

Email-to-Case

Inbound emails to a designated support address (support@yourcompany.com) automatically create Case records. Two flavors exist: On-Demand Email-to-Case (cloud-only, recommended) and Email-to-Case (requires a customer-hosted Email Agent service for inbound polling — rare in modern implementations).

Web-to-Case

A generated HTML form embedded on the customer’s website posts directly to Salesforce and creates a Case. Each org gets up to 5,000 Web-to-Case submissions per day. It’s the oldest channel and still useful for low-volume B2B sites, but most modern implementations replace it with Experience Cloud forms or messaging.

Messaging for In-App and Web (MIAW)

The current generation of chat/messaging. Replaced the legacy Live Agent product. MIAW supports embedded chat on web and mobile apps, persistent conversations, file attachments, and routing through Enhanced Omni-Channel. Sessions can spawn Cases automatically through an Omni-Channel Flow.

Service Cloud Voice

Voice integrates a contact center — Amazon Connect or partner telephony — directly into Salesforce. Voice calls are routed by Omni-Channel, transcribed in real time, and the resulting VoiceCall records can spawn Cases. Conversation Intelligence analyzes the transcript for keywords and surfaces Next Best Actions to the rep.

Experience Cloud (formerly Communities)

A self-service portal where customers can view their cases, file new ones via a contact request, search Knowledge, and interact with peers. The “Help Center” template is the current recommended starting point for an Experience Cloud-based service site.

Social Customer Service

Salesforce retired the original Social Customer Service feature, but the equivalent capability lives in Digital Engagement add-ons that connect Facebook Messenger, WhatsApp, SMS, and other channels through MIAW.

Omni-Channel Routing: Standard vs Enhanced

Omni-Channel is the routing engine that decides which service rep — or AI agent — receives an incoming work item. It supports cases, leads, custom objects, chats, voice calls, and messaging sessions. Crucially, Salesforce maintains two generations of Omni-Channel in parallel:

The four routing methods

Inside Omni-Channel, work can be routed using one of four configured methods. They’re not mutually exclusive — many orgs use multiple in parallel:

1. Basic Routing — Routing Configuration sends all work from a queue using a single routing model (Least Active, Most Available, External Routing). Simple and fast to configure.
2. Advanced Routing with Omni-Channel Flows — Routing logic built in Flow Builder using the Route Work action. Supports complex conditional routing, screen pops, and skill-based decisions.
3. Skills-Based Routing Rules — Map field values on the work item (case priority, account tier, language) to required skills, and route to the rep best matched. Configured from the Routing Configuration record.
4. Omni-Channel Unified Routing — A unified routing model for Voice scenarios that combines call queue and Omni-Channel queue behavior.

Routing Destinations: From Queues to AI Agents

Omni-Channel supports five distinct routing destinations. Each represents a different “what receives this work item?” answer. Knowing which destinations are supported in Standard vs Enhanced Omni-Channel is heavily tested:

Queue-based routing is conceptually the simplest: an admin creates a Queue with a list of members, attaches a Routing Configuration that picks the routing model, and Omni-Channel does the rest. Skills-based routing is more nuanced because it depends on a complete skill matrix — every rep must have the right skills assigned, and every work item must have its required skills set, or the work falls through to a fallback queue.

Case Automation: Rules and Order of Execution

Service Cloud has three case-specific automation rule types that predate Flow. They still run during the standard order of execution, and many production orgs rely on them every day:

When do the rules execute?

Case rules execute as part of the standard Salesforce order of execution, but the position matters. Assignment Rules fire before save (the case owner is set before before-save Flows run). Auto-Response Rules fire during save. Escalation Rules are separate — they’re evaluated by a time-based scheduler after the case has aged into the rule’s criteria.

Beyond the legacy rules, modern orgs increasingly use Flow Builder. Workflow Rules and Process Builder reached end of support on December 31, 2025 — Salesforce no longer accepts bug fixes or feature requests against them. Existing rules continue to execute, but all new automation must be built in Flow.

Einstein Case Routing

Einstein Case Routing is the AI-powered alternative to manual rule-writing. It analyzes historical case data and predicts the best queue, owner, or skill for incoming cases. Setup requires the Service Cloud Einstein feature license and at least a few thousand historical cases for the model to train on. Once active, Einstein routing can be invoked from an Omni-Channel Flow or run automatically at case creation.

Knowledge Management for Faster Resolution

Salesforce Knowledge is the article and FAQ system embedded inside Service Cloud. Service reps search it from the case page; customers browse it from Experience Cloud sites; and Agentforce Service Agents query it as their primary source of truth. Three flavors coexist:

Beyond Lightning Knowledge, Salesforce introduced Enterprise Knowledge (generally available) — a next-generation knowledge platform that adds Agentic Enterprise Search (AES) and integrates with Data 360 to unify external knowledge sources. Unified Knowledge can pull articles from third-party systems (SharePoint, Confluence, Zendesk) through Data 360 connectors, making them searchable inside Salesforce without migration.

Article lifecycle

1. Draft — Authors create articles in Lightning Knowledge with rich text, smart links, and embedded images.
2. Review & Approve — Optional flow-based approval process before publication.
3. Publish — Article becomes visible to assigned channels (Internal, Customer, Partner, Public).
4. Translate — Multi-language support via export/import workflow or in-platform translation.
5. Archive — Outdated articles are archived (not deleted) for compliance.

Knowledge visibility on the case page is controlled by Suggested Articles and Data Category Mapping. Suggested Articles uses the case Subject and Description to surface relevant content. Data Category Mapping pre-filters articles by case field values so reps see only what’s relevant to that case’s category.

The Service Console: Where Service Reps Work

The Service Console is the Lightning App that service reps live in. It’s a workspace-style UI — multiple cases open as workspace tabs, related records as subtabs, all with persistent utility bars at the bottom. Out of the box, the Service Console app includes:

• The Omni-Channel utility — accepts new work items, shows current capacity, tracks presence status
• The History utility — recent cases and records for quick recall
• The Notes utility — quick scratch notes attached to whatever record is open
• The Macros and Quick Text utilities — for repeatable actions and canned responses
• The Knowledge component — search and insert articles into emails and cases
• Case Feed — chronological activity timeline per case

Productivity features inside the console

Macros automate multi-step actions — send an email, change status, add a comment, all in one click. Quick Text provides reusable response snippets that can be inserted into emails and comments by typing a shortcut. Lightning Email Templates standardize outbound communications with merge fields and HTML formatting.

The Service Console also surfaces Record Alerts — contextual notifications that appear on a case or account when something needs the rep’s attention (e.g., “VIP customer,” “Open billing dispute”). Alerts are configured via Setup and integrate with Flow for dynamic triggers.

Entitlements, Milestones, and SLA Management

Service contracts in Salesforce are modeled with Entitlements and Service Contracts. An Entitlement defines the level of support a customer is entitled to (e.g., “24/7 Premium Support”) and links to an Account, Contact, or Asset. A Case inherits its entitlement from the related account, and the entitlement controls which Milestones apply.

Milestones

A Milestone is a time-based commitment inside an entitlement — “First Response within 1 hour,” “Resolution within 24 hours.” Milestones display on the case page with a countdown timer and trigger workflow when completed or violated. They’re the mechanical implementation of an SLA.

Milestones can be associated with Business Hours so they pause overnight, on weekends, or during configured holidays. This is how SLAs avoid penalizing the service team for time when no one was on duty.

Incident Management and Customer Service Swarming

When a single issue affects many customers — a service outage, a recalled product, a payment processor failure — handling each case in isolation is wasteful. Customer Service Incident Management addresses this with three new objects:

The natural flow is Case → Incident → Problem → Change Request. A customer reports an issue (Case). The support team realizes other customers are reporting the same thing and creates an Incident, linking the cases. Engineering investigates and identifies the underlying Problem. A Change Request is filed to fix it. Throughout, Broadcast Communications push status updates to every linked customer.

Customer service swarming

Swarming is Salesforce’s framework for cross-functional collaboration on high-severity issues. When a swarm is started on a case or incident, designated experts from other departments — engineering, finance, ops — are pulled in to collaborate in real time, typically through a Slack channel. Swarming integrates with Slack for the communication layer and with Salesforce for the record context.

Agentforce Service Agents and the AI Layer

The reason “Service Cloud” was renamed “Agentforce Service” is the dramatic expansion of the AI layer over the past two releases. AI in Service Cloud now spans three roles:

AI for the service rep (assistive)

Einstein for Service includes case classification, work summaries, reply recommendations, and article recommendations. Service Replies drafts an email reply for the rep based on case context. Work Summaries auto-generate case summaries when a case is closed. These are assistive features — the rep stays in control and edits the output.

AI for the customer (deflection)

Enhanced Bots are scripted conversational flows that handle simple customer questions before a human is engaged. Bots run on dialogues and intents and can escalate to a service rep when they can’t resolve the issue.

AI as the service rep (autonomous)

An Agentforce Service Agent is an autonomous AI agent that can resolve cases without human intervention. It uses generative AI grounded in Knowledge articles, plus configured Topics and Actions that let it update records, send emails, or trigger flows. Agentforce Service Agents are a first-class routing destination in Enhanced Omni-Channel — they appear alongside human reps in the queue and pick up work items just like a person would.

Agentforce Service Agents also work alongside the new Agentforce Contact Center, an all-in-one contact-center add-on generally available to Agentforce Service customers in the U.S. and Canada. It bundles voice, AI, and CRM into a single platform, eliminating the need for separate telephony integrations.

Supervising the Floor: Command Center for Service

The supervisor experience used to be called “Omni Supervisor.” Salesforce renamed it Command Center for Service and expanded what it shows. Command Center is the live operational view of everything happening in Service Cloud:

Supervisors can reassign work manually, change a rep’s queue or skill membership on the fly, whisper to a rep during a call, and pull historical reports — all from Command Center. With the latest Agentforce integration, supervisors can also instruct an AI agent to update a rep’s assignments using natural language (“Move the Spanish-speaking reps to the billing queue for the next two hours”).

Frequently asked questions

Wrapping up

Service Cloud (now Agentforce Service) is no longer just a case management product. The Case object is still the spine, but the body around it has grown — Omni-Channel handles every channel, Knowledge powers both reps and AI, Incident Management coordinates disruptions, and Agentforce Service Agents are quietly resolving cases at scale. Mastering the product means understanding how all of these layers cooperate, where the dividing lines between Standard and Enhanced sit, and which capabilities require which licenses.

For practitioners, that means staying current with at least the last three releases. For certification candidates, it means recognizing that the Salesforce certification track tests both the long-established mechanics (Cases, rules, entitlements) and the newer architecture (Enhanced Omni-Channel, Agentforce). You can browse all Salesforce certifications on CertifySF to find the exam that maps to your role.