Voice Automation Outcome Scorecard
Separate verified automated resolution, human-assisted resolution, unresolved containment and abandonment instead of treating every answered call as success.
Published 6 September 2026 · Original planning resources, not product certification or measured customer results.
What this tool measures
An agent can keep a caller on the line without resolving the request. This worksheet makes that distinction visible. Classify each included call once using your own verified evidence, then calculate outcome shares. The result reflects your supplied counts, not an independent audit of VoiceAvo.
Enter your planning inputs
Numeric defaults are fictional examples, not prices or measured performance. All calculations happen in your browser; inputs are not transmitted to the server.
Formula
Total classified calls = all four categories added together. Verified resolution (%) = (automatically resolved + resolved with human help) ÷ total × 100. Each other percentage uses its category count ÷ total × 100.
Worked example
For 60 automatically resolved, 20 human-assisted, 10 unresolved-contained and 10 abandoned calls, the classified total is 100. Overall verified resolution is 80%, while verified automated resolution is 60%. These percentages are deliberately not called simple containment.
Limits and assumptions
The four groups must be mutually exclusive and cover the included sample. Choose a consistent resolution definition and evidence window. Calls still pending or not reviewed should be excluded and reported separately rather than forced into a success category. Zero total produces an undefined rate, not 0% evidence of failure.
How to review the result
Check booking confirmations, completed tickets or other source-system evidence. Track transfers that connect but do not resolve separately in your working records. Compare like-for-like intent mixes and include unresolved cases when evaluating operational quality.
Exported JSON includes the inputs and derived outputs, making it possible for another reviewer to reproduce the arithmetic. Keep any underlying call evidence in your own approved system. A model result is not evidence that a live deployment has passed its acceptance tests.
Continue your evaluation
All six planning tools →Industry-specific call workflows →AI voice and IVR evaluation →Confirm the actual deployment
Provider support, integration behaviour and business rules must be checked for your own configuration. These materials do not establish a partnership, universal compatibility or a guaranteed outcome.
Explore VoiceAvo · Implementation checklist