How we measure the results we publish
The outcome figures on orvera.ai are averages from Orvera AI's own customer engagements. They are measured results from live deployments, not projections, and they are averages, so a single account can land above or below them. Once a use case is scoped, each account is modeled on its own numbers.
The figures, what each one means and how it is measured
| Figure | What it means | How it is measured |
|---|---|---|
| First-contact resolution of 90% | The share of conversations resolved on the first contact | A conversation counts as resolved when the customer makes no repeat contact about the same issue, on any channel, within 7 days. AI agents and human agents are counted together. |
| AI accuracy of 98% | The share of AI agent answers that are correct | AI agent answers scored correct in QA review against the customer's knowledge and policies, divided by all AI agent answers reviewed. |
| CSAT of 90% | The share of customers who rate the conversation well | Post-conversation survey answers of 4 or 5 on a 5-point scale, divided by all answers. |
| Cost per contact 25 to 30% lower | How much less each customer contact costs to handle | Total cost of handling contacts divided by the contacts handled, compared with the account's own baseline before go-live. |
| Average handle time 30% lower | How much shorter the conversations people handle get with Agent Assist | Each account's average handle time after go-live, compared with its own baseline before go-live. |
| 100% of conversations quality-scored | Every conversation is scored, human-handled and AI-handled, on every channel | AI Quality Management scores each conversation against the customer's own scorecard. |
| Full deployment in 3 to 6 weeks | Time from kickoff to the AI agents handling live traffic | Calendar weeks from kickoff to go-live on the first use case. |
Case study figures are different. Each one is the result of a single named or anonymized customer engagement, measured in that engagement, and it is reported with the case study it comes from.
Last updated