Agentforce for Service
Agentforce for Service is the Salesforce Agentforce SKU that ships two pre-built agents grounded in Service Cloud data: the customer-facing Service Agent and the rep-facing Service Coach.
Definition
Agentforce for Service is the Salesforce Agentforce SKU that ships two pre-built agents grounded in Service Cloud data: the customer-facing Service Agent and the rep-facing Service Coach. The Service Agent answers customer questions on Messaging, Web Chat, Slack, WhatsApp, and SMS, handling FAQs, order lookups, status checks, and basic record updates without a human in the loop. The Service Coach sits beside a live agent and offers reply drafts, case summaries, Knowledge suggestions, and next-step recommendations.
Both agents run on the Atlas Reasoning Engine and use Agent Actions that read and write Case, Contact, Asset, Knowledge Article, Order, and the various messaging objects. They ship with standard topics admins customize per industry: Order Status, Returns, Account Updates, FAQ Lookup, Escalation. Agentforce for Service is licensed per conversation. The customer-facing Service Agent's conversations are the primary cost driver in most deployments.
Why one SKU ships a customer-facing bot and a rep-facing coach
The Service Agent and the Service Coach in one SKU
The SKU bundles two agents that look almost opposite at runtime. The Service Agent is customer-facing, public-internet-exposed, and stateless across sessions (each conversation starts fresh). The Service Coach is rep-facing, internal-only, and stateful within an open case. The Service Agent's job is deflection: answer the question without a human if the answer is in Knowledge or in the customer's record. The Coach's job is augmentation: speed up the rep when deflection failed and the case opened. Together they form a tiered service model where the bot handles the routine and the coach makes the rep faster on the rest.
Channels the Service Agent can run on
Out of the box the Service Agent deploys to Salesforce Messaging (in-app, web embedded), Web Chat, WhatsApp Business, Apple Messages for Business, Facebook Messenger, SMS, and Slack. Each channel uses a deployment configuration that controls greeting text, transfer rules, and hours. The same agent serves all channels with topic and action sharing. Channel-specific behavior is handled via topic instructions that test which channel the session is on. Voice is a separate path; Service Cloud Voice integrates the agent as a real-time co-pilot for the human rep rather than running as a standalone voice bot, because voice bot quality has not caught up to text yet.
Pre-built topics and what they cover
The Service Agent ships with Order Status, Returns and Refunds, Account Updates, Shipping Address Change, FAQ Lookup, and Escalation. The Service Coach ships with Case Summary, Reply Drafting, Knowledge Suggestion, and Next Best Action. Each topic has a classification description and three to six attached actions. Most teams keep four or five Service Agent topics close to default and customize Returns and Refunds heavily because that is where company-specific policy lives. The Reply Drafting topic on the Coach is usually customized for brand voice the same way Email Drafting is on Sales.
Knowledge integration and the deflection metric
The FAQ Lookup topic queries Salesforce Knowledge through a Data Library or directly via a Knowledge Search action. When the agent finds a relevant article, it summarizes the answer rather than dumping the article text, and offers the source link as a citation. The deflection metric measures conversations where the agent answered the question without escalating to a human. Most well-tuned Service Agent deployments hit 30 to 45 percent deflection on routine inquiry mixes. The number drops fast when Knowledge is sparse or out of date. Knowledge hygiene is the largest predictor of Service Agent ROI in practice.
Escalation, transfer, and the human handoff
The Escalation topic catches messages the agent cannot handle and routes the customer to a queue, an Omni-Channel routing path, or a specific agent. Transfer behavior is configured per channel. On Messaging, the conversation transcript carries to the rep so they pick up mid-conversation. On Voice, the agent generates a case summary that pops on the rep's screen when the call connects. The Service Agent never asks the customer to repeat information it already collected, which is a small detail with a large impact on customer satisfaction. Tuning escalation triggers is a permanent activity, not a one-time setup task.
The Service Coach inside an open case
The Coach surfaces in the Lightning Service Console as a panel beside the case feed. It reads the case description, the contact history, attached Knowledge articles, and prior similar cases. Reply Drafting lets the rep ask for a draft response in a chosen tone; the Coach generates and the rep edits. Case Summary collapses a 40-message thread into three paragraphs the rep can paste into the resolution notes. Next Best Action surfaces the immediate suggested step ("propose RMA", "request photo of damage"). The Coach is invoked per click, not autonomously, which makes its conversation costs more predictable than the customer-facing agent's.
Licensing, conversation pricing, and deflection economics
Conversations on the customer-facing Service Agent dominate the cost. A high-volume B2C deployment can run hundreds of thousands of conversations a month. List price has run around $2 per conversation in 2026 with volume tiers. The ROI math is straightforward: cost per conversation against fully loaded cost per human-handled case (typically $5 to $15 depending on geography and channel). At 30 percent deflection on a $10 per-case cost, the SKU pays back at any conversation price under $7, but the math gets sensitive when deflection drops below 20 percent. Service Coach conversations are smaller in volume and easier to budget.
How to roll out Agentforce for Service without breaking customer trust
The cardinal rule of customer-facing bot rollouts is "fail to a human cleanly." Customers tolerate a bot that punts. They do not tolerate a bot that loops, gaslights, or hallucinates a policy. The sequence below puts that rule first.
- Audit Knowledge before turning on the agent
Pull every Knowledge article in the public-facing channel. Archive articles older than 18 months unless they are timeless. Confirm article titles match how customers actually phrase the question. Without this step the agent confidently cites wrong answers from stale content.
- Enable Agentforce for Service and provision both agents
Setup, Einstein, Agentforce, Enable. Accept terms. Provision Service Agent and Service Coach from the gallery. Both land in Agent Builder as editable copies of shipped templates.
- Customize Returns and Refunds for your actual policy
The shipped topic uses generic e-commerce policy. Replace the instruction and the Knowledge Data Library scope with your company's actual return windows, restock fees, and exceptions. This topic is the single biggest source of customer complaints when shipped without customization.
- Wire escalation to your existing Omni-Channel queues
Open the Escalation topic. Point its Transfer to Agent action at the queue you currently use for chat. Do not invent a new queue; the rep team must see escalated conversations in the same place they see all other work.
- Soft-launch the Service Agent on one channel with a beta opt-in
Pick Web Chat or Messaging. Put a small "Try our AI assistant (beta)" link rather than replacing the default chat entry point. Watch deflection and customer satisfaction for two weeks before expanding.
- Roll out the Service Coach to one rep team in parallel
The Coach is lower risk because it never speaks to a customer directly. A pilot team of five reps with weekly retrospectives identifies prompt and topic refinements before broad rollout.
- Watch the Plan Trace and a weekly conversation review
Pull a random sample of 50 conversations a week, read them with the service ops lead, and adjust topics where the agent picks wrong. This work never ends. Budget for it permanently.
The list of channels (Messaging, Web Chat, WhatsApp, SMS, Slack, etc.) the Service Agent is exposed on. Most rollouts start with one.
Which Knowledge articles, data categories, and channels the agent can read. Drives the FAQ Lookup topic accuracy.
The queue or routing path that catches conversations the agent cannot handle. Must integrate with the team's existing Omni-Channel setup.
The Coach's reply draft prompt template. Highest-leverage customization on the rep-facing agent.
When the Service Agent is active vs falls straight through to a queue. Useful when the agent's quality is uneven and you want human coverage during peak.
- The Service Agent without a Knowledge audit confidently cites stale articles. Treat the audit as a hard prerequisite, not a nice to have.
- Conversation pricing on the customer-facing agent surprises orgs that did not estimate volume. A B2C site with 100k monthly chat sessions and 60 percent agent coverage is 60k conversations a month.
- Escalation that drops the transcript is a customer satisfaction killer. Verify the handoff carries history before any external rollout.
- The Service Coach is mostly inert without Knowledge and prior case data. Empty Knowledge means generic suggestions that reps stop using.
- Voice is not yet a standalone bot path. Service Cloud Voice integrates the Coach for human reps, but a fully autonomous voice agent is not a shipped configuration as of mid-2026.
Trust & references
Straight from the source - Salesforce's reference material on Agentforce for Service.
- Agentforce for ServiceSalesforce Help
- Service Agent SetupSalesforce Help
- Service CoachSalesforce Help
About the Author
Dipojjal Chakrabarti is a B2C Solution Architect with 29 Salesforce certifications and over 13 years in the Salesforce ecosystem. He runs salesforcedictionary.com to help admins, developers, architects, and cert/interview candidates sharpen their fundamentals. More about Dipojjal.
Test your knowledge
Q1. What does Agentforce for Service reason over when handling a customer inquiry?
Q2. How does Agentforce for Service handle situations it cannot resolve?
Q3. Across which channels can Agentforce for Service operate?
Discussion
Loading discussion…