Now Advisory · Buyer side guide · 2026 edition
Now Assist Usage Monitoring: A Buyer Side Guide
What to monitor across assists, actions and overage, how to set internal thresholds, and how to catch overage before the top up charge, with benchmark data from real enterprise renewals.
Section 01Why now assist usage monitoring protects the budget
Now assist usage monitoring is the discipline that keeps a metered AI deployment inside its budget, because in a consumption model the bill is set by behaviour you can only manage if you can see it. Now Assist draws assists from a bundled allowance, and once that allowance is exhausted, overage triggers top up charges. Monitoring is how you watch consumption against the allowance in time to act, rather than discovering the overage on an invoice. This guide explains what to monitor and how, with benchmark data from real enterprise renewals.
We are independent advisors with no vendor partnership and nothing to resell. For the wider context, start with our overview of Now Assist pricing and our service on Now Assist consumption advisory.
Section 02What now assist usage monitoring means
Usage monitoring means tracking three things continuously: how many assists you are consuming, which workflows and users are consuming them, and how that consumption tracks against your bundled allowance over the period. It is not a one time check at renewal. It is an ongoing operational cadence that turns the metering data into a managed budget rather than a surprise.
The reason it matters is that consumption is not flat. It climbs with adoption, spikes with new use cases, and concentrates in a few heavy workflows. Without monitoring, the first visibility into any of this arrives with the overage charge. Our guide to the ServiceNow assist consumption model explains the metering you are monitoring.
Section 03What to monitor: assists, actions and overage
Three signals matter most. First, total assist consumption against the bundled allowance, expressed as a percentage of the period's envelope so you can see the trajectory. Second, consumption by workflow and action type, because large agentic actions consume materially more than simple ones and a few heavy workflows often drive most of the burn. Third, the pace toward overage, which is the rate at which you are approaching the allowance, not just the current total.
Watching all three together tells you not only whether you will exceed the allowance but when, and which workflow is responsible. That attribution is what lets you act on the cause rather than just observe the cost. Our Now Assist consumption model guide sets out how the heavy actions stack up.
Section 04Setting internal thresholds
Monitoring is only useful if it triggers action, and that requires thresholds set in advance. Define consumption checkpoints across the period, for example alerts when you reach defined fractions of the bundled allowance ahead of schedule, so a fast burn is flagged while there is still time to respond. The thresholds should reflect your own tolerance, not wait for the vendor to notify you of overage.
Set thresholds at the workflow level too, so a single runaway agent is caught before it consumes a disproportionate share of the allowance. The point of a threshold is to convert a silent trend into a prompt to act, well before the allowance is exhausted and the top up rate engages.
Section 05Reading the metering data
The metering data tells a story if you read it correctly. A steady climb in line with planned adoption is healthy. A sudden step change usually means a new use case went live or an agent is running more often than intended. A concentration of consumption in one workflow signals where to focus governance. Reading these patterns is the difference between data you collect and data you act on.
The discipline is to review the data on a regular cadence rather than only when something breaks. Trends are visible early if you look; they are expensive if you wait. An experienced reading of the metering data turns it from a record of what happened into a forecast of what is about to.
Section 06Catching overage before the top up charge
The whole purpose of monitoring is to catch overage before it becomes a charge. Overage is not a single event; it is the end of a trajectory you can see building if you are watching the pace toward the allowance. When the trajectory shows you will exhaust the envelope before the period ends, you have choices: throttle heavy workflows, defer non essential usage, or open a renegotiation for additional bundled capacity at the committed rate rather than the top up rate.
Acting early is almost always cheaper than acting late. A renegotiation conducted before you are in overage is a negotiation; one conducted after the top up charges have started is damage control. Monitoring is what buys you the time to be in the first situation rather than the second.
Section 07Monitoring as a renewal input
Monitoring data is one of the most valuable inputs to the next renewal. A clear record of actual consumption, by workflow and over time, tells you precisely what allowance you need and where the vendor's default falls short. It also gives you evidence that the vendor cannot dispute, because it is your own usage data, which is the strongest possible anchor for negotiating allowance and overage terms.
Without monitoring, you renew on the vendor's assumptions about your usage. With it, you renew on facts. Feed the consumption history into the renewal model and use it to size the next allowance precisely, the way our Now Assist pricing overview describes.
Section 08Building the monitoring cadence
Monitoring works as a cadence, not a tool you check occasionally. Establish a regular review of consumption against allowance, with defined owners, defined thresholds, and a defined response when a threshold is crossed. The cadence should be frequent enough to catch a fast burn within the period, which for most enterprise rollouts means at least monthly and often more during a ramp up.
Build the cadence into operations from day one of the rollout, not after the first overage. The teams that control assist cost are the ones that treat monitoring as a standing process with clear ownership, the same way they would treat any other consumption based spend. An independent advisor can help design the cadence and the thresholds against benchmark consumption.
Section 09Monitoring mistakes to avoid
The recurring mistakes are clear. Treating monitoring as a renewal time check rather than an ongoing cadence. Tracking total consumption but not attribution, so you see the cost but not the cause. Setting no thresholds, so the data informs nobody until overage arrives. And failing to feed the consumption history into the next renewal, so you negotiate the allowance on assumptions instead of facts.
Each is avoidable. Monitor continuously, attribute consumption by workflow, set thresholds that trigger action, and carry the history into renewal. Done that way, usage monitoring keeps a metered deployment predictable and turns your own consumption data into leverage at the table.
Section 10An illustrative monitoring scenario
Consider a deployment three months into an enterprise rollout where assist consumption has been climbing faster than planned. Without monitoring, the first signal would be the overage charge at period end. With monitoring in place, a threshold alert fires when consumption reaches a defined fraction of the allowance ahead of schedule, and the attribution data points to a single agentic workflow running far more often than its design intended.
That early signal changes the options entirely. The team can throttle the runaway workflow, correct the configuration, and, if the trajectory still points to overage, open a renegotiation for additional bundled capacity at the committed rate well before the period closes. The same situation without monitoring ends in a top up charge and a harder conversation. These dynamics reflect typical benchmark observations rather than a specific account.
Section 11Who owns Now Assist usage monitoring
Monitoring only works when it has an owner. Assign clear responsibility for reviewing consumption against allowance on a defined cadence, for acting when a threshold is crossed, and for carrying the consumption history into the next renewal. Without named ownership the data accumulates unread and the first person to notice the trend is the one approving the overage invoice.
In most enterprises the ownership sits across ITAM, the platform team, and procurement, which is why the cadence and the thresholds need to be agreed rather than assumed. An independent advisor can help design the monitoring process, set thresholds against benchmark consumption, and ensure the resulting data feeds the renewal as evidence rather than being lost between teams.
FAQFrequently asked questions
What is Now Assist usage monitoring?
It is the ongoing discipline of tracking how many assists you consume, which workflows and users consume them, and how that tracks against your bundled allowance over time. Because Now Assist is metered, monitoring is how you manage the budget and catch overage before it becomes a top up charge.
What should we monitor in Now Assist?
Three signals: total assist consumption against the bundled allowance, consumption by workflow and action type since large agentic actions consume materially more, and the pace toward overage. Watching all three tells you whether you will exceed the allowance, when, and which workflow is responsible.
How do we catch overage before it is charged?
Set internal thresholds that alert when you reach defined fractions of the allowance ahead of schedule, review the metering data on a regular cadence, and act on the trajectory: throttle heavy workflows or open a renegotiation for additional bundled capacity at the committed rate before the top up rate engages.
How does monitoring help at renewal?
Monitoring produces a record of actual consumption by workflow over time, which tells you precisely what allowance you need and gives you usage data the vendor cannot dispute. That evidence is the strongest anchor for negotiating allowance size and overage terms at the next renewal.