Onsi Labs · Calculator

Cost-to-Serve Calculator

An indicative comparison of your current assisted-servicing cost against an AI-assistant deployment, using published channel-cost benchmarks. Three inputs; adjust the assumptions to match your book.

3 inputsInstant resultsFree — no sign-up

Your book

Total members across the schemes you service.
Assisted contacts only — calls, chats, emails handled by a person.
Gartner's cross-industry median is $13.50. Use your own fully loaded figure if you have it.
Display only — enter the cost figure in your currency.
60%
The share of contacts an AI assistant resolves without a person. Deployed virtual-agent studies report ~64% as a well-implemented baseline; routine benefits queries (coverage, claims status, eligibility) contain at the top of the range, complex claims near zero. Below ~30%, the case weakens — treat with caution.

Executive summary

Indicative annual saving
$1.05m
52% of current servicing cost, before platform and implementation costs
Servicing cost today
$2.02m
150,000 assisted contacts / year
With an AI assistant
$976k
60,000 contacts still handled by people

Assumptions in play: contained interactions move to self-service-class cost ($1.84 per contact, scaled from Gartner's $1.84 median); the remaining 40% stay at your assisted cost of $13.50.

Test this case honestly. Below roughly 30% containment, or on books under ~50,000 lives, the economics weaken — the whitepaper's framework treats these inputs as a signal to model category by category before committing, not as a case to buy on a blended promise.

Built on industry research
GartnerQuarterly Journal of EconomicsMcKinseyIBM
Updated July 2026

Model it properly, on your book

The four variables that determine any deployment's economics — volume and mix, containment by category, fully loaded channel costs, and the revenue side — are empirical, not assumed. Onsi helps insurers and benefit providers model them honestly, with the failure modes priced in, and deploy AI assistants within a governed, auditable framework.

Indicative only. Figures use published third-party benchmarks — Gartner's median cost per contact ($13.50 assisted, $1.84 self-service) and deployed virtual-agent containment baselines — applied to the inputs you provide. Results are not a representation about outcomes on any particular book of business, exclude platform and implementation costs, and exclude agent-productivity gains on residual contacts (peer-reviewed research indicates +14% on average). The full framework, sources and caveats are set out in the whitepaper.

Built by Onsi LabsTrusted AI for insurance, benefits and wellbeing.Copyright © Onsi.