Now Advisory · Buyer side guide · 2026 edition
Now Assist pilot pricing: a buyer side guide
Now Assist pilot pricing is where production cost is either understood or set up to surprise you. This guide shows how to run a pilot that measures assist weight, not just value, and lock production terms before you commit, with benchmark data from real enterprise renewals.
Section 01What a Now Assist pilot should price
Now Assist pilot pricing should answer one question: what will this cost at production scale? A pilot that only proves the AI works has done half the job, because in the 2026 model AI is metered in assists and the cost is a consumption question. The pilot's most valuable output is not a value story; it is a measurement of how many assists each workflow draws, which is the number that determines production cost when the same workflow runs at full volume.
The buyer side framing is that a pilot is a measurement instrument, not a sales proof. The vendor runs pilots to demonstrate value and build momentum toward a commitment; the buyer should run the same pilot to extract the consumption data that prices production honestly. Used this way, Now Assist pilot pricing protects the production decision rather than rushing it, because the commitment that follows is sized from measured assist draw rather than optimism.
This guide sits under the Now Assist pricing pillar and pairs with our Now Assist consumption advisory. It treats the pilot as the cheapest place to learn what production will cost.
Section 02The pilot to production trap
The pilot to production trap is the gap between a low volume pilot and full production, and it is where Now Assist pricing most often goes wrong. A pilot runs a few workflows for a small population at modest volume, drawing few assists, so the consumption looks comfortable. Production runs many workflows for the whole estate at full volume, drawing many assists, and the comfortable pilot number does not scale linearly because agentic adoption deepens as well as broadens.
The trap closes when a commitment is sized on pilot consumption and then exhausted by production within the term, surfacing as overage at the overage rate. The pilot was never wrong; it was simply measured at the wrong scale and projected as if scale did not change the mix. Avoiding the trap means treating the pilot as a per action measurement, not a total volume forecast. Our Now Assist overage guide details how this gap becomes a top up charge.
Production assist consumption is frequently several times pilot consumption, driven by both wider adoption and deeper agentic use. A commitment sized on pilot volumes is the most common route into mid term overage.
Section 03Measuring assist weight in the pilot
The pilot's job is to measure assist weight per action, the number of assists a given workflow draws each time it runs, with particular attention to agentic actions. A simple generative request draws roughly one assist; an agentic action that plans, retrieves, calls tools and loops draws many. The pilot is the place to observe these weights directly, because it is small enough to instrument and real enough to be representative.
The measurement to capture is consumption per workflow class, not aggregate pilot consumption. Aggregate consumption tells you what the pilot drew; per class weight tells you what production will draw when each workflow scales, which is the number that prices the commitment. A pilot that records only its total has measured the wrong thing for pricing purposes. The method connects to our Now Assist consumption advisory.
Instrument the pilot to report assist draw by workflow from the start, because retrofitting the measurement after the pilot ends loses exactly the data that prices production.
Section 04Pricing production from pilot data
Production pricing combines two numbers: the assist weight per workflow class, which the pilot measures, and the production volume per workflow class, which the rollout plan estimates. Multiplying weight by volume across all workflows that will adopt gives modelled production consumption, which prices against the committed pool and overage rate. This separates what the pilot can teach, per action cost, from what only the rollout plan can supply, total volume.
Modelling production this way produces a range rather than a point, because volume estimates carry uncertainty. The range tells finance the likely production cost and the worst plausible case, and it sizes the committed pool and overage contingency. It also reveals how sensitive production cost is to agentic mix, which is the variable that most often surprises. The budgeting side is developed in our Now Assist budgeting guide.
Price production as pilot measured weight times planned volume, expressed as a range. A single projected number from pilot totals is the trap; a weighted model is the way out of it.
Section 05Terms to lock before the pilot ends
The leverage to set production terms is highest before the pilot converts to a commitment, because until you commit, the alternative of not proceeding is live. Once production workflows depend on the AI, that alternative is gone and the terms are whatever the contract left open. The discipline is to negotiate the production terms while the pilot is still a pilot, using its data as evidence rather than waiting until adoption has removed your leverage.
The terms to lock are the overage rate, the committed pool sized to modelled production, rollover so a conservative pool is protected, documented agentic weighting so consumption is calculable, and a mid term resize right so the commitment can grow with adoption. Securing these before the pilot ends turns the production decision into a planned step rather than a commitment made under the momentum the pilot built. Our Now Assist negotiation guide sets out the agenda.
A pilot that ends without production terms agreed has handed the vendor the moment of maximum buyer commitment and minimum buyer leverage, which is exactly the moment to have settled the terms.
Section 06Negotiating from pilot evidence
Pilot data is negotiation evidence the vendor cannot easily dispute, because it is their own platform's measurement of your own usage. A buyer who arrives at the production conversation with measured assist weights and a modelled production range sets the terms of the discussion; a buyer who arrives with only a successful pilot story accepts the vendor's sizing and rate. The evidence shifts the conversation from whether the AI works to what its consumption honestly costs.
This is also why the pilot should be instrumented from the buyer side, not just observed through the vendor's success framing. The same pilot can produce a value narrative for the vendor and a consumption model for the buyer, and the buyer needs both to negotiate from strength. The leverage dynamics are the same as any renewal, concentrated at the pilot to production boundary. Our Now Assist pricing pillar carries the wider picture.
Negotiated from evidence, the production commitment is sized to reality; negotiated from momentum, it is sized to the vendor's hopes.
Section 07The Now Assist pilot pricing ask
A disciplined pilot converts into a short set of production terms, each grounded in measured pilot data.
- Measured assist weights
Consumption per workflow class observed in the pilot, instrumented from the start.
- Modelled production range
Pilot weights times planned volumes, expressed as a range with an agentic mix sensitivity.
- Right sized pool
A committed pool matched to modelled production, not pilot totals.
- Fixed overage rate and rollover
A stated overage rate and rollover secured before the pilot converts to commitment.
- Documented weighting and resize right
Agentic weighting written in and a mid term resize right so the commitment grows with adoption.
Presented together while the pilot is still open, these size the production commitment from evidence and lock the terms while leverage is highest.
Section 08The pilot pricing checklist
Before a Now Assist pilot converts to a production commitment, confirm each item below from pilot measurement rather than the vendor's success story.
- Pilot instrumented to report assist draw per workflow class from the start.
- Assist weight of agentic actions measured directly in the pilot.
- Production consumption modelled as weight times planned volume, as a range.
- Committed pool sized to modelled production, not pilot totals.
- Overage rate fixed, rollover secured and agentic weighting documented before commitment.
- A mid term resize right in place so the commitment grows with adoption.
If the pilot ends with a value story but no consumption model, the production commitment will be sized on optimism and corrected by overage. Measure the weight, model the range and lock the terms before you commit.
FAQFrequently asked questions
What should a Now Assist pilot price?
A Now Assist pilot should price production cost, not just prove the AI works. Because AI is metered in assists, the pilot's most valuable output is the assist weight per workflow, the number that determines production cost when the same workflow runs at full volume. Treat the pilot as a measurement instrument, not a sales proof.
What is the pilot to production trap?
It is the gap between a low volume pilot that draws few assists and full production that draws many. Consumption does not scale linearly because agentic adoption deepens as well as broadens, so a commitment sized on pilot consumption is often exhausted by production within the term and surfaces as overage.
How do we price production from pilot data?
Combine the assist weight per workflow class measured in the pilot with the production volume per class from the rollout plan. Weight times volume across adopting workflows gives modelled production consumption, expressed as a range, which sizes the committed pool and the overage contingency.
Which terms should we lock before the pilot ends?
Fix the overage rate, size the committed pool to modelled production, secure rollover, document agentic weighting and add a mid term resize right, all before the pilot converts to a commitment. Leverage is highest while not proceeding is still a live alternative.
Are these figures official ServiceNow prices?
No. All ranges are typical negotiated figures based on benchmark observations across real enterprise renewals, used as internal leverage rather than published as official list prices.