Loading page
About this tool
Pick containment, CSAT and deflection metrics before launch and project deflected volume and savings.
This planner fixes the definition of every chatbot success metric — containment, escalation, deflection, CSAT, goal completion and fallback rate — before the bot goes live, and projects deflected contact volume as monthly conversations multiplied by the target containment rate. It is for support leads and conversation designers who have to defend a bot's business case, and who need each metric's formula agreed with analytics before the first conversation is logged rather than argued about after launch.
Open Chatbot Analytics Metric Planner on AltFTool — it loads instantly in your browser.
Add your input to the workspace.
Adjust the options until the result looks right.
Copy or download the output and put it to work.
Every metric ships with the exact numerator and denominator to instrument.
Each metric names the way it is most commonly gamed or misread.
Turns a containment target into deflected contacts and an annual saving.
It depends entirely on scope: a bot answering a narrow set of FAQ intents can reasonably contain 60-80%, while one fronting complex account or billing work often sits at 20-40%. Compare against your own pre-launch baseline rather than a published benchmark, because containment is defined differently by almost every vendor.
Containment rate = conversations the bot resolved with no human handoff divided by total bot conversations, times 100. The critical decision is whether abandoned conversations count as contained — they should not, so require an explicit resolution signal such as a confirmed answer, a completed task or a no-return window before counting one.
Containment measures what share of conversations the bot finished on its own; deflection measures how many human contacts were actually avoided. They differ because some bot conversations would never have become a ticket, so deflection is always the smaller and harder number, and savings math built on raw containment overstates the result.
At minimum containment, escalation, goal or task completion per flow, fallback rate and bot CSAT with its response rate. Instrument all of them before the first release — a metric added three months later has no comparable history, so you cannot tell whether a change helped.