How we measure and label outcomes
Every performance figure on this site links here. This page defines exactly how each metric is calculated, over what period, for which customers, and how we label results so nothing is overstated.
What each metric means
Resolution rate
Share of incoming support interactions Auralis resolves end to end without a human, measured over a trailing 30-day period once live.
Deflection
Interactions fully handled in self-service or by the agent so they never reach a human queue. Reported separately from resolution.
Productivity
Change in human handling time per interaction after Auralis Agent Assist, versus the 30 days before go-live.
Cost-per-ticket reduction
Change in fully-loaded cost to serve per interaction, using your own labor and tooling costs, over the measurement period.
Measurement period
A trailing 30 days, reported no earlier than 30 days after go-live so figures reflect steady state, not launch week.
Cohort rules
Cohort averages include only production dealers with at least 30 days live and 1,000+ interactions in the period.
Exclusions
We exclude pilot and sandbox traffic, test tickets, and the first 14 days after go-live (ramp).
How results are labeled
Every figure is labeled Customer-cohort average, Highest observed result, Expected range, or Customer-specific result — and customer-specific results carry the customer and date.
TRANSPARENCY
Want the numbers for your dealership?
We'll model expected outcomes against a sample of your own e-automate ticket mix, using these definitions.