Resolution as a Service (RaaS) is the pricing and architectural model in which enterprise software is priced on problems solved rather than the number of users who log in. That definition contains a word that the current wave of SaaS pricing transitions is systematically skipping: architectural.
ServiceNow’s Q1 2026 earnings call produced the most consequential pricing disclosure in enterprise software this year. Bill McDermott confirmed that 50% of the company’s net new ACV now comes from non-seat-based pricing. Renewal rates held at 97%. Guidance was raised. The stock still fell 17 to 18% on April 23, the largest single-day decline in the company’s history.
The market is not confused. The market is reading ahead.
The 50% Figure Is the Right Data Point and the Wrong Conclusion
The bulls are correct that ServiceNow is the most advanced public incumbent on the transition spectrum. The 50% non-seat ACV number is the strongest single data point available for anyone arguing that the migration away from headcount-linked pricing is real and executable at scale. CPAG has cited it as exactly that.
But the bulls are drawing the wrong conclusion from it. The argument that NOW has “already addressed” the bear thesis rests on treating non-seat-based revenue as equivalent to resolution-based revenue. It is not.
ServiceNow’s non-seat ACV almost certainly includes token-based Now Assist consumption, volumetric workflow automation fees, and platform activity charges tied to the Action Fabric. That is a consumption architecture. The billing unit changed from seats to tokens and workflow triggers. The accountability model did not change.
This is the Results-as-a-Service failure mode. Results as a Service frames the AI pricing transition as an outcome billing problem: change the invoice structure and the transition is complete. Resolution as a Service frames it as a structural architecture problem. The distinction is not semantic. It is the dividing line between vendors who can defend their revenue at renewal and vendors who discover too late that they changed the contract before they changed anything that mattered.
What the Four-Layer RaaS Stack Actually Requires
ServiceNow’s Action Fabric and agentic connectivity work maps cleanly to Layer 3 of the Four-Layer RaaS Stack. That layer is real and valuable. But Layer 3 alone is not a RaaS architecture.
Layer 2 is the High-Fidelity Repository: the graph-structured institutional knowledge architecture that gives AI agents the domain context, audit trails, and attribution records needed to execute and prove resolutions at scale. It is not a database. It is the institutional logic layer that connects what happened to why it was resolved, and generates the evidentiary record that answers the customer’s question at renewal: how many problems did you actually solve, how do you know, and how do I verify it?
There is no public evidence that ServiceNow has built or is building an attribution-grade measurement architecture of this kind, one specifically designed for independently verifiable resolution accounting at the contract level. There is no indication that Atomic Resolution criteria, meaning work that is verifiable (a problem solved, not attempted), attributable (traceable to the platform’s AI execution, not a concurrent human process), and finite (a clear endpoint preventing open-ended agentic loops), are embedded in their contract structure or their measurement infrastructure. The Action Fabric handles agentic connectivity. It does not handle attribution.
Billing on tokens consumed or workflows triggered is not billing on problems verified as solved. The Friction Economy does not shrink just because the billing unit changed. The customer’s CFO, watching AI-driven efficiency metrics and preparing for next renewal, will eventually ask a single question: what did you resolve, and can you prove it? Token consumption data does not answer that question.
The Ghost Seat Problem Has a Longer Fuse Than Renewal Rates Show
High renewal rates and strong RPO growth are backward-looking metrics. They measure stickiness in existing contracts. They do not measure forward resistance to SaaS Demotion.
SaaS Demotion is the dynamic by which MCP-enabled AI agents bypass vendor UIs entirely, eliminating the interface layer where traditional vendors capture value. If ServiceNow’s enterprise customers begin routing agentic workflows through Microsoft Copilot orchestration or Salesforce’s agent layer rather than the ServiceNow UI, the renewal rate will lag the revenue risk by 12 to 24 months. Ghost Seats, licensed seats paid for but no longer actively used because AI absorbed the workflow, appear in the data at renewal, not before.
The capital markets are not making a mistake about ServiceNow’s fundamentals. They are repricing the probability that its current architecture survives the next three years of agentic maturity.
This is the correct read. A vendor billing on token consumption can face its own version of the Churn Cascade if general-purpose agent infrastructure absorbs the workflow routing function it currently monetizes. The exposure has changed shape. It has not been eliminated.
What a Complete Transition Actually Looks Like
The Three-Phase RaaS Transition Roadmap defines the gap between where ServiceNow is and where a complete transition requires going.
Phase 1, the Revenue Audit (months 1 to 6), maps ARR exposure to seat risk and identifies Ghost Seat rates by customer segment. This is the diagnostic that establishes which renewal conversations become difficult within 12 to 18 months regardless of what the vendor does.
Phase 2, the Hybrid Pilot (months 6 to 18), instruments the platform to track resolutions per customer and cost-to-serve per resolution type against existing seat and consumption contracts, before any customer is converted to outcome pricing. This is where the High-Fidelity Repository build begins in earnest. The goal is to establish Resolution Contribution Margin benchmarks by resolution type in a controlled cohort before any customer is converted to outcome pricing.
Phase 3, Full RaaS Diversification (years 3 to 5), converts customers to outcome-based pricing using the evidence base Phase 2 produced. The target is seats below 20% of revenue, Resolution Contribution Margin by resolution type validated at institutional-grade gross margin thresholds, and the Repository compounded into an un-rippable moat.
ServiceNow’s current position looks like a company partway through Phase 2, with the commercial transition partially underway but the measurement trust infrastructure and attribution layer still incomplete. That is more progress than most incumbents can credibly claim. It is also not enough to answer what capital markets are actually asking.
The Valuation Debate Is the Wrong Frame
The appropriate CPAG response to the “big mispricing” bull argument is not to debate whether NOW is a good or bad stock. It is to reframe the question.
The valuation debate treats ServiceNow’s transition as a binary: either they’ve made it or they haven’t. The RaaS architecture argument treats it as a spectrum with a specific missing component. The bulls are correct that NOW is less exposed than a pure seat vendor. They are wrong that “non-seat ACV” and “resolution-based revenue” are the same thing, and they are wrong that the transition is structurally complete.
The question that determines ServiceNow’s 2030 revenue trajectory is not whether 50% of their new ACV is non-seat-based. It is whether they can price on verified resolutions, not workflow activity, before their enterprise customers’ AI maturity outpaces the attribution capabilities of their measurement infrastructure.
That question has not been answered. The 50% figure, as significant as it is, does not answer it.
The philosophical dimension of what it means for an institution to price accountability rather than access, and what measurement infrastructure makes that accountability real rather than contractual fiction, is explored in depth at Middle Way in AI. The ServiceNow case is the clearest current illustration of why that question is not abstract.
Prescription
Run the attribution test against your own pricing architecture before your customers run it against yours. Identify your three highest-volume workflow categories. For each one, answer three questions: can you prove the problem was solved and not merely attempted, can you isolate your platform’s execution from concurrent human or external processes, and does the resolution have a defined endpoint you can document in a billing record? If any answer is no, you are running a consumption model with outcome language attached to it. That is not a RaaS architecture. It is Business Model Debt with a new invoice format.
What would a CFO at your largest customer find if they ran that same test on your last renewal invoice?