ETHORITY®CUSTOMER SERVICE · SCENARIOS
What could a supervised AI service change?
Customers ask AI directly. People supervise quality and handle exceptions. Explore how useful capacity, adoption and full costs change the case.
Three fictional scenarios share one baseline. “Typical” is a planning hypothesis, not a measured industry average. Net time released already deducts supervision, knowledge upkeep and rework.
Change the assumptions
Potential capacity value after costs
- Current annual staff-effort equivalent
- €600,000
- Useful capacity-value equivalent before AI and setup costs
- €240,000
- Year-one capacity-value equivalent after costs
- €180,000
- Share of current staff-effort baseline
- 30 %
Preset scenarios: best case €180,000 / 30%; typical planning case €60,000 / 10%; failed pilot −€30,000, its full pilot budget. These are potential capacity equivalents, not promised ROI or cash savings.
Why the best case needs deliberate work
ETHORITY can help define the routine, permitted knowledge, human authority and escalation, then test actual adoption, accepted outcomes and usable capacity. New customer behaviour may require another channel mix or workflow. Expansion follows evidence; the percentage is not a delivery commitment.
Calculation: requests × minutes ÷ 60 × hourly cost × released share × useful share − annual AI running cost − one-time setup. Cash savings require observed budget or spend changes. Taxes, financing and unquantified service harm are excluded; the pilot budget does not cap all possible business losses.
Inspect the capacity-conversion method →Advanced: test review time for AI-assisted drafts
This separate response-level model assumes people approve every draft. It preserves the earlier assumptions and sensitivity controls; its baseline differs from the direct-support scenario above.
CUSTOMER SERVICE · DECISION MODEL
What changes when checking takes longer?
Adjust the service example’s assumptions. Count review, rework and costs before interpreting released capacity.
This is a hypothetical sensitivity model, not a customer result or provider quotation. No preset is an observed baseline. Inputs remain in this browser; the enquiry continuation stores its context locally and prepares an unsent draft.
Inspect and change the assumptions
Modelled result under these assumptions
- Assisted attempts per year
- 48,000
- Accepted responses per year
- 47,040
- Net capacity after review and rework
- 3,568 h/year
- Annual capacity-value equivalent after operating cost
- €51,336
- Year-one capacity-value equivalent after operating and setup costs
- €11,336
Potential capacity is not cash savings. The conversion share is an assumption about destination work; attribution and realised financial value require their own observations. Negative results remain visible.
Method and source
The response-level scenario model of 14 Oct 2026 calculates attempts from annual volume × eligibility × assistance share. Accepted responses apply the acceptance share. Net hours subtract review on every attempt and rework on unsuccessful attempts from the minutes released on accepted responses. Positive hours apply the assumed destination-work share and hourly cost; negative hours retain their full cost. Annual operating and one-time setup costs are then deducted.