Bot Performance
Bot Performance is the Salesforce surface that reports on how an Einstein Bot or Agentforce Bot is doing across the metrics that matter operationally: total sessions, average session length, intent recognition rate, containment rate (the percentage of conversations the bot resolved without escalation), transfer rate, customer satisfaction, and per-dialog success metrics.
Definition
Bot Performance is the Salesforce surface that reports on how an Einstein Bot or Agentforce Bot is doing across the metrics that matter operationally: total sessions, average session length, intent recognition rate, containment rate (the percentage of conversations the bot resolved without escalation), transfer rate, customer satisfaction, and per-dialog success metrics. The page lives inside Einstein Bot Builder under the Performance tab and complements custom CRM Analytics dashboards built on the underlying Bot session objects. Bot Performance is the dashboard sales operations and customer experience teams open every week to track whether the bot is paying back its investment.
Bot Performance matters because bot value is invisible without measurement. A bot that resolves 80 percent of inbound chats saves significant agent time but only if the team can prove it; a bot that resolves 10 percent and frustrates customers needs urgent attention but only if the data surfaces the problem. The Performance tab provides the headline metrics; CRM Analytics or custom Lightning dashboards extend them with cohort analysis, intent-by-intent breakdowns, and longer-term trends. Mature bot programs treat Bot Performance as a weekly operational meeting input and tune the bot continuously based on what the metrics reveal.
The metrics that actually matter for bot performance
Containment rate
Containment is the percentage of bot sessions resolved without escalating to a human agent. A 70 percent containment rate means seven of every ten conversations end inside the bot. This is the single most-watched bot metric because it ties directly to operational savings. High containment plus high CSAT signals a healthy bot; high containment with low CSAT signals deflection rather than resolution.
Intent recognition rate
The percentage of user utterances the bot's NLP model correctly classified into an intent. A low rate (under 70 percent) means users hit Sorry I didnt understand responses too often and abandon the bot. Improving recognition usually means adding utterances to existing intents or splitting an overloaded intent into more focused ones.
Transfer to Agent rate
The complement to containment. Tracks how often a session ends with a transfer. Some transfers are intentional design (escalations on confirmed complex cases); others are failure modes (intent miss, action failure, dead-end dialog). The page typically breaks the transfer count by trigger so teams can distinguish the two.
Average session length
Sessions that complete in two or three turns are usually successful self-service. Sessions over ten turns often indicate a stuck conversation or repeated clarification loops. The metric is informative when paired with containment; long contained sessions are positive, long transferred sessions are negative.
Customer satisfaction (CSAT)
Post-conversation CSAT surveys ask the user to rate the bot. The aggregate score is the gold standard for bot health. CSAT below the human-agent baseline signals that automation is hurting the experience even if it saves money; CSAT at or above the baseline justifies expansion.
Per-dialog success rate
The Performance tab surfaces per-dialog metrics: how often each dialog is entered, how many users complete it, and where they drop off. The metric is the foundation for tuning specific dialogs; the dialog with the highest drop-off rate is the highest-leverage improvement target.
Custom dashboards beyond the Performance tab
The native page covers headlines. Custom CRM Analytics dashboards extend coverage with cohort analysis, intent-by-intent trend, message-by-message drilldown, and integration with Knowledge attachment effectiveness. Most mature bot programs run both.
Common pitfalls
Three patterns recur. Watching containment without CSAT confuses deflection for resolution. Acting only on aggregate metrics misses dialog-specific drop-offs. And ignoring intent recognition issues leaves users hitting Sorry I didnt understand replies that quietly destroy trust. Each is addressable with disciplined weekly review.
How to operate Bot Performance weekly
Bot Performance is most useful as a recurring operational habit, not a one-off audit. Establish the cadence early and tune continuously.
- Open the Performance tab
Einstein Bot Builder, open the bot, click Performance. The page shows containment, intent recognition, transfer rate, and per-dialog metrics for the selected date range.
- Pair every metric with CSAT
Containment, intent recognition, and transfer rate are operational metrics. CSAT is the customer signal. Always read them together to avoid optimising for the wrong outcome.
- Investigate per-dialog drop-offs
Find the dialog with the highest drop-off rate. Open it in Bot Builder, walk the steps, identify why users abandon. Fixes here have the highest leverage.
- Improve intent training when recognition lags
For intents with low recognition, add utterances. Real customer phrasings come from transcripts of failed conversations; mine those for new training data.
- Hold a weekly review
Treat Bot Performance as a weekly operational meeting. Without cadence, the data accumulates without action.
- Containment without CSAT confuses deflection for resolution. Always read the two together.
- Aggregate metrics hide dialog-specific drop-offs. Drill into per-dialog data weekly.
- Intent recognition issues quietly destroy user trust. Sorry I didnt understand replies are the loudest signal of a training gap.
- Performance data is retained for a limited window. Snapshot to a custom history object for trend analysis beyond a few weeks.
Trust & references
Cross-checked against the following references.
- Bot PerformanceSalesforce Help
- Einstein Bots OverviewSalesforce Help
Straight from the source - Salesforce's reference material on Bot Performance.
- Bot MetricsSalesforce 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 Bot Performance track?
Q2. Why is a high transfer-to-agent rate on a specific dialog a red flag?
Q3. Why pair CSAT with resolution rate when evaluating bot performance?
Discussion
Loading discussion…