If you searched for Salesforce Service Cloud and landed here, you’re not lost. Salesforce renamed the product in October 2025 at Dreamforce, and folded in a set of AI features that go well past a name change. The console your agents live in, the data model underneath it, and the AI layer running on top of it all shifted at once.
This guide breaks down what Agentforce Service actually is, what’s new since the Service Cloud days, and what to check before you plan a migration or a fresh rollout.
What Is Agentforce Service? (Quick Answer)
Agentforce Service is Salesforce’s AI-powered customer service platform, built on the Salesforce Platform, Agentforce, and Data Cloud. It’s the product formerly known as Service Cloud. It still manages cases, routing, knowledge, and channels the way Service Cloud did, but it now ships with AI agents that can resolve routine cases on their own, draft SLA updates, and hand off complex cases to human reps with full context attached.
Salesforce reports that Agentforce Service offloads up to 72% of routine inquiries to AI agents, which contributes to 32% faster case resolution and an estimated 33% lift in CSAT scores (Salesforce FY26 customer success metrics).
How Agentforce Service Works
Three pieces do the heavy lifting:
- The Salesforce Platform — the CRM foundation: cases, accounts, contacts, permissions, and everything that made Service Cloud reliable in the first place.
- Agentforce — the AI agent layer. This is what drafts responses, recommends next steps, and can act autonomously within guardrails you set.
- Data Cloud — the data layer that grounds those agents in real, current customer information instead of stale records or guesswork.
Without Data Cloud feeding it clean, unified data, an AI agent is just guessing with confidence. That’s why most Agentforce Service rollouts start with a Data Cloud implementation before anyone touches agent configuration, the agent is only as reliable as the data underneath it.
Core Features of Agentforce Service
Case Management, Now With Agentic Milestones
Case management still works the way it always did: every inquiry from email, chat, phone, or social becomes a case with a full interaction history attached. What’s new is Agentic Milestones, introduced in the Summer ’26 release. You can mark specific SLA milestones as agentic, and Agentforce drafts and sends the routine updates, “we’re still working on this”, on its own, based on rules you define in Prompt Builder. Reps stop typing the hundredth status update and start spending that time on cases that need judgment.
Command Center for Supervisors
Supervisors get a real-time view across service, sales, and marketing touchpoints, not just service cases. They can watch live conversations across channels, step in when a human needs to take over, and correct an AI agent’s guidance mid-conversation through Slack if it’s giving customers outdated information. In one example Salesforce cites, catching a misconfigured agent early enough this way saved 26% in support costs before the issue snowballed, worth noting as a single case, not a guaranteed outcome.
Unified Service Rep Workspace
This replaces the old multi-tab Service Console experience. Reps see purchase history, past cases, product usage, and even a customer’s mood, inferred from language and tone, in one screen, with inline suggestions that update as the conversation moves. Less searching, less guessing, faster resolution.
Agentforce Service Portal (formerly part of Experience Cloud)
Self-service moved from static FAQ pages to a portal that answers natural-language questions and can walk a customer through a multi-step task, like updating a shipping address mid-return. It also flags issues before a customer notices them — low product usage, an upcoming renewal and reaches out first. If you’re trying to cut ticket volume through case deflection, this is the piece to get right; we’ve written a deeper breakdown of how to design a case deflection portal if you want the specifics.
Voice, Messaging, and Channel Consolidation
Service Cloud Voice is now Agentforce Voice, and it added real-time voice-to-text switching in Enhanced Chat v2, along with custom chat window styling and delivery receipts. Worth flagging for anyone still on standard channels: Salesforce is retiring Standard Facebook Messenger and Standard SMS channels in mid-2026, so if you haven’t migrated to Enhanced channels, that move is overdue rather than optional.
MCP for Agentforce
Model Context Protocol support means Agentforce agents can now connect to external systems and AI models through an open standard, rather than needing a custom integration built for each one. If your service agents need to check inventory in an ERP system or pull data from a tool outside Salesforce, MCP is what makes that a configuration task instead of a development project. We cover this pattern from the sales side in our piece on turning Slack into an AI revenue operations assistant, and the same connective logic applies to service.
Refined Agent Analytics
Service Agent Analytics and Employee Agent Analytics merged into one dashboard with more than 40 metrics, so you can measure human reps and AI agents side by side instead of pulling two separate reports and reconciling them by hand. Custom Scorers (beta) let you grade agent sessions against your own KPIs rather than only Salesforce’s default quality metrics.
Agentforce Service by the Numbers
- 72% of routine inquiries can be offloaded to AI agents (Salesforce FY26 metrics)
- 32% faster case resolution with unified context in the rep workspace
- ~33% estimated CSAT improvement from proactive, personalized outreach
- 65% of service leaders expect case volume to rise this year, versus 58% who expect headcount to grow
- 85% of service decision-makers expect service to contribute a larger share of company revenue
- Service budgets are projected to grow by an average of 23% over the next year
(Sources: Salesforce Agentforce Service announcement, Salesforce State of Service report)
On the customer side, independent research keeps landing in similar territory: PwC and Qualtrics XM Institute both report that somewhere between 72% and 86% of consumers say they’d pay more for a better service experience, depending on the study and year. The exact number moves, but the direction hasn’t changed in years: poor service has a price tag, and it’s getting more expensive.
The Data Model Behind Agentforce Service
If you’re new to the platform, or you’re a Service Cloud admin catching up, the core objects haven’t changed much; they’re just doing more now that AI agents interact with them directly.
| Object | What it represents |
|---|---|
| Account | A company or organization in your system |
| Contact | An individual linked to that account |
| Case | The customer issue or request being resolved |
| Case Comment | Internal notes added to a case |
| Case History | A log of every update made to a case |
| Knowledge | Your library of support articles and documentation |
| Entitlement | The support level a given customer is entitled to |
| Milestone | SLA deadlines tracked against a case, now support agentic automation |
| Queue | Distributes cases by priority and workload |
| Macro | Automates repetitive sequences in the console |
Agentforce adds a parallel layer on top of this: Agent, Topic, and Action records that define what an AI agent is allowed to do and how it reasons through a case, but the underlying case data model is the same one Service Cloud admins have worked with for years.
Agentforce Service vs. Agentforce Sales: What’s the Difference
Sales Cloud picked up the same treatment and is now Agentforce Sales, so it’s worth a quick distinction since the two are often confused inside the same org.
Agentforce Sales manages pipeline, opportunities, forecasting, and everything involved in closing a deal, now with an AI-powered Sales Workspace and qualification agents built in.
Agentforce Service manages cases, support workflows, and everything involved in keeping a customer satisfied after that deal closes.
Most organizations run both side by side. Agentforce Sales owns the relationship up to purchase; Agentforce Service owns it from there. If your team is weighing which one to prioritize first, our Sales Cloud implementation page covers the sales side in more depth.
Cloud Odyssey’s Point of View
We’ve walked a fair number of clients through this exact transition, and the pattern repeats: teams get excited about the AI agent capability and skip the groundwork that makes it work. Agentic case handling and proactive outreach are only as good as the data behind them. If your Account and Contact records are duplicated, your case data is inconsistent, or your knowledge base hasn’t been touched since the pandemic, an AI agent trained on that data will confidently give a customer the wrong answer, just faster than a human would have.
Our recommendation for anyone moving from Service Cloud to Agentforce Service: start with a Salesforce CRM health check before turning on agentic features. Fix the data and process gaps first, then layer in AI agents where they’ll actually reduce case volume instead of just automating the same mess faster. We saw this play out with CurrencyFair, where cleaning up the query-handling process before deploying Agentforce is what made the automation stick.
If you’re planning that move, whether it’s a first-time Salesforce implementation or an upgrade from an existing Service Cloud org, our Salesforce Service Cloud team can walk through what’s genuinely useful for your case volume versus what’s still marketing. And once it’s live, our managed services team keeps it tuned as Salesforce keeps shipping new releases, which, based on the last twelve months, isn’t slowing down.
Talk to our Salesforce Service Cloud team →
FAQs About Agentforce Service
Salesforce folded its entire “Cloud” portfolio under the Agentforce brand at Dreamforce ’25, Sales Cloud became Agentforce Sales, Marketing Cloud became Agentforce Marketing, and Service Cloud became Agentforce Service. The name signals where new investment is going: AI agents that resolve cases, draft SLA updates, and act on customer data as a built-in part of the product rather than an add-on. It lines up with Salesforce’s own research showing 65% of service leaders expect case volume to rise while only 58% expect headcount to keep pace, the rebrand is Salesforce’s answer to that gap, not just a marketing refresh.
Service Cloud hasn’t disappeared, it’s the same product under a new name, now with an AI layer on top. Existing orgs keep their configuration, data, and automations exactly as they were; the rename alone doesn’t force a migration. Upgrading to the newer agentic features, though, does require the right license and some setup work.
You can run core case management without it, but the AI features, agentic case handling, proactive outreach, and personalized recommendations depend on unified, accurate customer data to work reliably. Salesforce is in the process of rebranding Data Cloud to Data 360, so you may see either name depending on where you’re reading. Running Agentforce without a clean data foundation tends to produce agents that sound confident and guess wrong.
No. Agentforce is the broader AI agent platform that spans sales, service, marketing, and more. Agentforce Service is the service-specific product built on top of it, alongside Data Cloud (Data 360) and the core Salesforce Platform.
