Case study · Insurance · Now Assist cost control
This ServiceNow insurance now assist cost control case study shows how an insurer forecast its Now Assist consumption, right sized the committed pool, fixed the overage rate, and held assist spend within around 5 percent of forecast using buyer side analysis and benchmark data from real enterprise renewals.
Of additional assist cost avoided versus open ended pricing
Variance of first year assist spend to forecast
Weeks before the deadline the renewal closed
A national insurer reached its ServiceNow renewal with the account team proposing a substantial Now Assist expansion alongside the standard uplift. The pitch framed a large committed assist pool as the efficient way to adopt the bundled AI across claims and service workflows. The insurer brought us in buyer side to forecast what its assist consumption would actually be before committing to a pool sized by the vendor, using benchmark data from real enterprise renewals.
In the 2026 model, AI is bundled across Foundation, Advanced and Prime and assists are metered, so the cost of Now Assist is driven by consumption rather than a flat fee. Large agentic actions draw materially more assists than routine ones, which makes the committed pool a forecasting problem, not a list price problem. The insurer had no internal model of how many assists its claims and service workflows would consume, so the vendor's proposed pool was the only number on the table.
Our first task was to replace that number with a forecast the insurer owned. A committed assist pool sized above genuine demand becomes shelfware that cannot be recovered, while a pool sized below demand exposes the insurer to overage top up charges. The right pool sits at realistic consumption with margin, and finding it required modelling rather than accepting the proposal.
A consumption forecast across the insurer's intended use cases produced a materially smaller pool than the vendor proposed. A large share of the planned workflows used routine assists that consume little, while the smaller set of agentic actions that consume heavily was concentrated in a few high value claims processes. Sized to that pattern, the committed pool needed to be well below the proposal, with a defined growth allowance for adoption rather than a large fixed commitment paid from day one.
The forecast also exposed the overage exposure hidden in the proposal. Without a fixed overage rate, any consumption beyond the pool would have been priced at the vendor's discretion, turning the unpredictability of agentic actions into open ended cost.
We built the response around the forecast. First, we sized the committed assist pool to modelled consumption with margin, removing the unrecoverable shelfware the larger pool would have created. Second, we negotiated a fixed overage rate so any consumption beyond the pool was priced at a known, discounted number rather than at the vendor's discretion. Third, we secured a growth allowance on assists so adoption could scale into the buffer before any overage applied.
The insurer's procurement and IT asset teams led every session. We stayed behind the table, building the consumption model, drafting the evidenced counters and briefing executives, in the pattern set out in our ServiceNow renewal negotiation advisory and the wider ServiceNow renewal guidance.
The renewal closed four weeks before the deadline. The right sized pool and the growth allowance removed the unrecoverable commitment in the original proposal, and the fixed overage rate closed the open ended exposure, avoiding what would have been roughly 40 percent of additional assist cost had consumption been priced at the vendor's discretion. Across the first year the insurer held its assist spend within around 5 percent of forecast. The approach mirrors the consumption discipline in our other anonymised engagements, including our media Now Assist cost control case study and our retail Now Assist cost control case study.
Three lessons carry beyond this engagement. A vendor proposed assist pool is a forecast the buyer has not made, and sizing the pool to your own modelled consumption is where most of the value sits. Because large agentic actions consume materially more than routine ones, the split between heavy and light use cases decides the right pool size, not the headcount. And a fixed overage rate paired with a growth allowance turns the unpredictability of metered assists into a known cost path rather than open ended exposure. Final contract language should be reviewed by counsel.
It describes an insurer facing a large vendor proposed Now Assist pool at renewal. A buyer side consumption forecast right sized the committed pool, secured a fixed overage rate and a growth allowance, and held assist spend within around 5 percent of forecast.
The case study is anonymised. It is based on real enterprise renewal engagements, with the client profile, estate and figures presented as plausible and internally consistent ranges rather than naming any organisation.
Consumption forecasting did most of the work. The vendor proposed pool was sized above genuine demand, so modelling the split between heavy agentic actions and routine assists produced a smaller, safer pool with a fixed overage rate and a growth allowance.
No. All figures are typical negotiated ranges based on benchmark observations across real enterprise renewals, used as internal leverage rather than published as official list prices.
Your renewal