Now Advisory · Buyer side guide · 2026 edition
Now Assist consumption forecasting that holds up at renewal
How to model assist demand across adoption scenarios, where vendor forecasts go wrong, and how a defensible Now Assist consumption forecast sizes the commitment correctly.
Section 01Why Now Assist consumption forecasting decides your bill
Now Assist consumption forecasting is the discipline of estimating how many assists your workloads will draw across the contract term, and it is the single most important input to your AI cost. Under the 2026 model, AI is bundled into every tier but assists are metered, so the forecast determines the size of the commitment you buy and the overage you risk. Get the forecast right and the commitment fits. Get it wrong and you either overcommit to unused capacity or undercommit into top up charges.
The reason forecasting matters more than it used to is that consumption is variable in a way seat counts never were. A licence is a known quantity for the year. Assist consumption shifts with adoption, workflow design and the growing share of agentic actions that draw materially more assists than simple ones. A forecast that ignores this variability produces a commitment that fits on paper and fails in practice. For the underlying model, see our Now Assist pricing guide.
The buyer side goal is a forecast you can defend with your own data, because the alternative is accepting the vendor's forecast, which is built to size the commitment to their revenue target rather than your demand. Forecasting is leverage, not just budgeting.
Section 02Building the consumption baseline
Every forecast starts from a baseline, and the best baseline is pilot data. If you have run Now Assist on a subset of users or workflows, the observed assist draw per user and per workflow is your most reliable evidence. Extrapolate from real consumption rather than from a generic per user assumption, because the generic figure rarely matches how your teams actually use the platform.
If you have no pilot, build the baseline from the workflows you intend to enable. Estimate the share of simple actions versus large agentic actions for each workflow, weight the agentic actions for their higher assist draw, and multiply by expected volume. The weighting is the part most forecasts miss, and it is the part that matters most, because agentic actions are where consumption concentrates. Our Now Assist budgeting guidance walks through translating the baseline into a budget figure.
Whichever route you take, instrument your usage so the baseline improves over time. The detail on tracking actual draw against forecast sits in our Now Assist usage monitoring guidance, and the feedback loop it creates is what keeps the forecast honest across the term.
Section 03Modelling adoption scenarios
A single point forecast is fragile because adoption is uncertain. The buyer side approach is to model a range: a conservative scenario where AI is enabled on a subset of workflows, an expected scenario reflecting your realistic rollout plan, and an aggressive scenario where adoption runs ahead of plan and agentic actions take a larger share. The spread between them is the information you need.
The spread tells you two things. First, how much headroom to negotiate into the commitment, because a wide spread means more risk of crossing the allowance under faster adoption. Second, how much the overage rate matters, because if the aggressive scenario sits well above the committed pool, the rate above it becomes a material exposure rather than a footnote.
Size the committed pool to the expected scenario with enough headroom to absorb the gap toward the aggressive case, rather than sizing to either extreme. Sizing to the aggressive case overcommits you to capacity you may never use. Sizing to the conservative case guarantees overage the moment adoption succeeds. The expected case with headroom is the defensible middle.
Section 04Where vendor forecasts go wrong
Vendor consumption forecasts go wrong in predictable directions, and recognising the pattern is half the defence. The most common is an optimistic adoption assumption that sizes the commitment high, locking you into capacity you may not reach. The second is a conservative pool paired with a punitive overage rate, which looks cheap at signature and generates top up charges as you ramp. Both serve the vendor, in opposite ways.
The third failure is treating the assist draw per action as flat, ignoring the difference between simple and agentic actions. A forecast that does not weight agentic actions will understate consumption for any estate moving toward autonomous workflows, which is most of them. The result is a commitment that fits today's usage and fails tomorrow's.
The independent move is to bring your own forecast, built from your data and weighted for action type, and to compare it openly against the vendor's. The negotiation then becomes a reconciliation of two models rather than an acceptance of one, which is exactly where a buyer with evidence wins.
Section 05Benchmark observations on forecasting accuracy
Based on benchmark observations across real enterprise renewals, buyers who forecast from pilot data and weight for agentic actions size their committed pool closer to actual consumption than buyers who accept a vendor estimate. The accuracy gain shows up as fewer surprise top up charges in the second and third years, which is where consumption pricing usually bites.
The forecast also strengthens every other term. A defensible consumption number supports the case for a larger pool at the committed rate, a fixed overage rate and a conversion right, because each ask is backed by evidence rather than assertion. Forecasting is not only budgeting, it is the foundation of the negotiating position.
These outcomes are typical negotiated results used as internal leverage, not official list prices. The right forecast depends entirely on your adoption profile, which is why it has to be built from your own usage.
Section 06Turning the forecast into renewal terms
A forecast only earns its value when it becomes contract terms. Translate the expected scenario into a committed pool sized with headroom, the spread toward the aggressive case into the overage rate you fight for, and the uncertainty in the model into a conversion right that lets sustained overage become an expansion at the committed rate. The forecast is the evidence behind each of these asks.
The forecast also informs term length. If your adoption path is steep and uncertain, a shorter commitment or a mid term review point may protect you better than locking a multi year pool to an early estimate. If adoption is predictable, a longer commitment at a fixed rate may be worth the certainty. The forecast tells you which.
If your renewal includes Now Assist and you want a defensible consumption forecast built before the vendor sets the commitment, book a renewal assessment call and we will model the demand from your own usage data.
Section 07Keeping the forecast honest after signature
A forecast is a living instrument, not a one time exercise. Once the commitment is signed, the forecast should be checked against actual consumption every period so that drift is caught early. If real usage is running ahead of the expected scenario, you have time to act before the allowance is exhausted, whether by re scoping an inefficient workflow or triggering a conversion at the committed rate. If usage is running behind, you carry that evidence into the next renewal to argue against an inflated commitment.
The discipline pays off most at the next renewal, when the vendor returns with a fresh estimate. A buyer who has tracked actual draw against forecast across the term arrives with the strongest possible evidence: their own measured consumption. That record replaces the vendor spreadsheet as the reference point and turns the renewal into a reconciliation of reality rather than a negotiation over assumptions. Building and maintaining that model is the core of our Now Assist consumption advisory service.
The feedback loop also sharpens the workflow design that drives consumption in the first place. A forecast that surfaces which workflows draw the most assists tells you where efficiency work has the highest return, which protects the commitment without sacrificing the automation outcomes. Forecasting, monitoring and design improvement reinforce each other across the term, and the buyers who treat them as one discipline keep their AI cost predictable in a way that point estimates never achieve.
Forecasting is ultimately a leverage discipline as much as a budgeting one. A buyer who walks into the renewal with a defensible model built from real usage negotiates from evidence, while a buyer who relies on the vendor estimate negotiates from assertion. The model supports every ask, the committed pool size, the overage rate, the conversion right and the term length, and it carries forward into the next renewal as the reference point. The work compounds, because each term of disciplined forecasting makes the next negotiation stronger and the cost more predictable
FAQFrequently asked questions
What is Now Assist consumption forecasting?
It is estimating how many assists your workloads will draw across the contract term. Because AI is bundled but assists are metered, the forecast determines the size of the commitment you buy and the overage you risk, making it the most important input to your AI cost.
How do I build a consumption forecast?
Start from a baseline, ideally pilot data, then weight for the mix of simple and large agentic actions. Model conservative, expected and aggressive adoption scenarios and use the spread to set the committed pool size and the headroom you negotiate.
Why not just use the vendor forecast?
Vendor forecasts tend to either size high to lock in capacity or size low with a punitive overage rate. Many also treat assist draw as flat, ignoring that agentic actions consume more. Bring your own weighted forecast and reconcile the two models.
How does the forecast affect renewal terms?
It is the evidence behind every ask: a committed pool sized with headroom, a fixed overage rate, a conversion right and the right term length. A defensible forecast turns each negotiating position from assertion into evidence.
Are these forecasting figures official ServiceNow prices?
No. They are typical negotiated outcomes based on benchmark observations across real enterprise renewals, used as internal leverage. They are not official list prices and depend on your adoption profile.