Calculate Cost per Resolved Enquiry for a Voice Assistant
Calculate cost per resolved enquiry with a transparent worked example that includes unresolved calls, human follow-up and clearly defined outcome evidence.
Table of Contents▼
Cost per resolved enquiry connects voice-assistant spending to a useful business result. It can reveal more than a low price per minute, because inexpensive conversations may still leave staff doing the work afterwards. The calculation is only trustworthy when “resolved” has a clear definition.
Start by choosing one enquiry type and one reporting period. Mixing opening-hours questions with complex service issues can hide meaningful differences in effort. A focused calculation helps an operator decide whether a particular workflow is improving, even before the business has enough data for a broader return-on-investment model.
Define resolution before counting it
For a simple information request, resolution might mean a supported answer was provided and the caller stated no further need. For a request requiring an action, require evidence that the permitted action reached its intended final state.
A call ending normally is not sufficient. Nor is a request merely being captured if the business still needs someone to resolve it. Use a separate pending category and keep uncertain cases visible.
Burki's ROI guide provides a broader framework. This narrower measure should feed that analysis rather than being presented as profit or proven labour savings on its own.
Choose the cost boundary
Include the costs needed to operate the workflow during the period: platform usage, separately billed providers, telephone charges and relevant fixed fees. Add the human follow-up effort you actually measure. Do not assume that every automated minute removes a paid minute of work.
Avoid double counting. If a platform charge includes a component, do not add an estimate for that same component again. If providers bill directly under BYO, include those invoices where relevant. Twilio's invoice guidance illustrates why a call-log price may omit other services.
State whether setup effort is included, amortised over an explicit period or reported separately. Each choice can be reasonable if the comparison uses the same boundary.
Work through a hypothetical example
Suppose an illustrative monthly pilot has 200 eligible enquiries. After review, 120 are resolved by the assistant, 50 require staff follow-up and 30 are incomplete or uncertain. Assume total workflow operating cost is 180 currency units. These numbers are fictional, not Burki prices or customer results.
The assistant-only cost per resolved enquiry is 180 divided by 120, or 1.50 units. Reporting 180 divided by 200 would instead describe cost per eligible enquiry, which is a different metric.
Now suppose staff spend a measured 100 minutes finishing the 50 follow-ups, valued for this hypothetical calculation at 0.40 units per minute. That adds 40 units. If all 50 are independently confirmed resolved, the combined process cost is 220 divided by 170, approximately 1.29 units per resolved enquiry.
The lower combined figure does not mean human work was free. It reflects a larger verified resolution denominator and an explicitly included follow-up cost.
Test the sensitive assumptions
Change the resolution count, follow-up time and setup allocation separately. Ask which assumption most affects the result. If a small relabelling of uncertain calls dramatically improves the metric, the outcome definition needs more scrutiny.
Compare like periods and enquiry mixes. A quiet week of easy questions should not be used to claim savings against a busy week of complex problems. Show the counts alongside the ratio so readers can see the underlying workload.
State how repeat enquiries are counted. If the same issue generates three calls before resolution, counting three resolved enquiries would overstate the result. A stable case identifier or a documented deduplication rule can help the business separate interaction volume from the number of problems actually resolved.
Use the result to choose an improvement
If follow-up effort dominates, improve the handoff information or narrow the assistant's scope. If many calls remain uncertain, improve outcome evidence before optimising price. If provider costs dominate, review the actual workload and billing categories.
Burki's cost tracking guide can support that next investigation. Calculate one workflow honestly, retain the unresolved cases and use the result to decide where operational effort will matter most.
Ready to try Burki?
Create an assistant and check your available browser practice allowance.
Start Free TrialTrial eligibility and available practice are shown in your workspace.