Chat
Chat in Salesforce is the embeddable web-chat channel that lets customers start a text conversation with a service agent (or an AI bot) from a website.
Definition
Chat in Salesforce is the embeddable web-chat channel that lets customers start a text conversation with a service agent (or an AI bot) from a website. It originated as Live Agent in early Service Cloud, was rebranded to Chat with the Service Cloud Lightning console, and now sits inside Salesforce Service Cloud Digital Engagement as one of several Messaging channels (alongside SMS, WhatsApp, Apple Messages for Business, Facebook Messenger). Chat conversations route through Omni-Channel to available agents and persist as Live Chat Transcript or Messaging Session records in Salesforce, complete with the full message history, customer info, and any attached Knowledge articles.
Chat matters because text-based support is now table stakes for most consumer-facing brands. Customers prefer chat to phone calls for many issue types; agents handle multiple chats concurrently with capacity controls, which makes Chat one of the most cost-effective service channels per resolved case. Salesforce's Chat product is the embedded widget plus the agent-side console plus the routing and analytics on top. Modern deployments increasingly front-load Chat with Agentforce or Einstein Bots to handle the long tail of common questions, escalating only to human agents when the bot cannot resolve. The result is dramatically higher containment (the percentage of conversations the bot handles end-to-end) with no loss of customer experience.
How Salesforce Chat fits modern service
The embedded Chat widget
Salesforce ships an embeddable Chat widget that drops into any website with a few lines of JavaScript. The widget shows a Chat Now button (or proactive prompt), opens a chat window when clicked, and connects the visitor to Omni-Channel for agent or bot routing. Visual customisation includes branding colours, button text, and pre-chat forms.
Agent surface and the Service Console
Agents see incoming chats inside the Service Console's Omni-Channel widget. Accepting a chat opens it in a primary tab with the customer's record details, the transcript so far, and a typing area. Agents can handle multiple chats concurrently up to their configured capacity (typically three to five).
Omni-Channel routing
Omni-Channel routes incoming chats based on agent availability, capacity, and skills. Routing configurations specify which agents handle which chat queues, with skill-based matching to send high-tier or specialist chats to the right agents. Without Omni-Channel, Chat falls back to manual queue handling, which scales poorly.
Bot front-end via Einstein Bots and Agentforce
Most production Chat deployments now route incoming conversations to an AI bot first. Einstein Bots handles the legacy product surface; Agentforce represents the modern reasoning-engine generation. The bot answers common questions, looks up account data, and resolves end-to-end where possible; escalation to human agents happens only when the bot cannot resolve.
Pre-chat forms and routing context
Pre-chat forms collect basic info (name, email, issue category) before the conversation starts. The data routes the chat to the right agent skill and prepopulates the case context. Most production deployments use pre-chat forms; skipping them produces agent-side cold starts that slow resolution.
Chat Transcript and persistent record
Salesforce stores each conversation as a Live Chat Transcript (legacy Chat) or Messaging Session (Digital Engagement) record with the full message log, agent who handled it, duration, and any attached articles. Reports on these objects answer chat volume, average handle time, and CSAT questions.
Reporting and analytics
Standard reports include chats by hour, handle time by agent, CSAT by skill, abandonment rate, and bot containment. Most service operations build dashboards combining these metrics; weekly review against KPIs is how chat programs improve over time.
Common pitfalls
Three patterns recur. Capacity caps set too high (six or seven concurrent chats) produce slow responses and customer frustration. Pre-chat forms with too many fields drop conversion before the chat even starts. And bot fallback paths that fail to recognise escalation requests produce dead-end interactions. Each is addressable with tuning and continuous measurement.
How to deploy Salesforce Chat
Standing up Chat is mostly Omni-Channel routing plus embedded-widget configuration. The discipline that makes it succeed is bot design, capacity tuning, and reporting cadence.
- Enable Chat (Live Agent or Digital Engagement)
Setup, Chat Settings. Configure deployments, chat buttons, and queue assignments. Service Cloud Voice and Digital Engagement editions ship with Messaging Sessions instead of Live Chat Transcripts.
- Configure Omni-Channel routing
Create Service Channels for Chat, Routing Configurations, and Presence Configurations. Tune capacity (three to five chats per agent is typical) and skill-based routing.
- Build the embedded widget
Configure the Chat embed snippet with branding, pre-chat form fields, and proactive prompt triggers. Drop the snippet into the website pages where chat should appear.
- Front-load with Agentforce or Einstein Bots
For high-volume programs, route incoming chats to a bot first. The bot handles common questions; humans handle the rest. Containment 70 percent or higher is achievable for mature bots.
- Build chat dashboards
Report on chat volume, handle time, CSAT, abandonment, and bot containment. Weekly review against KPIs is what improves chat programs over time.
- Capacity caps too high slow responses and frustrate customers. Three to five concurrent chats per agent is the practical sweet spot.
- Pre-chat forms with too many fields drop conversion. Ask for the minimum needed for routing; collect details after the chat starts.
- Bot fallback paths that miss escalation requests produce dead-end interactions. Test escalation explicitly.
- Live Chat (legacy) and Messaging Sessions (Digital Engagement) use different data models. Reports built on one do not work for the other.
Trust & references
Cross-checked against the following references.
- Salesforce ChatSalesforce Help
- Digital Engagement MessagingSalesforce Help
Straight from the source - Salesforce's reference material on Chat.
- Omni-Channel OverviewSalesforce 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 is Chat in Salesforce currently branded as?
Q2. How are Chat conversations routed to agents?
Q3. What can be combined with Chat to deflect simple inquiries?
Discussion
Loading discussion…