An answered call is an activity measure. A completed caller task is an outcome measure. Track both, and define exactly what counts before comparing a service with your current operation.
A proprietary simulated benchmark describes a different evidence set from your live pilot. Do not substitute it for your own outcome definitions. Loman’s company-published accuracy methodology ↗
If you report a completion rate, define which calls were eligible and how completion was verified. A wrong-number call should not be treated like a failed order. A reservation request outside the restaurant’s policy may be handled correctly without becoming a booking.
Keep excluded calls visible in a separate category. Removing difficult cases from the denominator can make a report look better while hiding the work your team still has to do. Document the rules and apply them consistently across service models.
For routine information, check that the answer matched the approved source. For a system action, verify the receiving record. For a handoff, confirm that an appropriate owner received the request. For an unresolved call, record why it remained open and what happens next.
This structure helps distinguish a successful boundary from a failure. An AI correctly declining to guess about an allergy and reaching a manager is different from giving an unsupported answer. A service taking a reservation request is different from confirming available inventory.
Sample ordinary calls as well as exceptions. Include busy and quiet periods, different call reasons and updates to restaurant information. Review privacy and recording requirements before collecting or sharing call material. Give reviewers a simple checklist so the result does not depend only on personal impressions.
Track repeat problems and their cause. A stale menu file calls for a source update, while a dead transfer may need a routing repair. Record when the fix was made and rerun the relevant scenario. Do not close a problem solely because someone changed a setting.
Use actual billed usage once available. Compare it with the quote assumptions and investigate differences in billable minutes, call volume or escalation work. Measure staff capacity separately from payroll changes so you do not overstate savings.
End each review with a decision: continue, change a defined part of the scope, retrain or adjust information, or stop a route that is not working. Assign each change an owner. A compact recurring review is more useful than a dashboard that shows many answered calls but no record of what the restaurant or guest actually received.
Answer rate tracks whether a call was answered. Completion rate tracks whether an eligible caller task reached the defined outcome. State the denominator, exclusions and verification method for each.
No. Reaching staff can be the correct outcome for a complaint or uncertain allergy question. Judge the call against the policy, and distinguish a completed handoff from a transfer attempt.
No. A call can stay within automation but still end with a wrong answer or missing order. Review verified task completion and unresolved exceptions alongside containment.
Editorial operating guidance from Loman. The checklists are planning aids, not independent performance research or legal advice. Confirm your restaurant’s current policies and provider support.
Official Loman product information ↗Use the same call tasks and acceptance criteria when you evaluate Loman.
Learn more about Loman ↗Featured entity