Joe Davis

The large majority of enterprise AI pilots fail for one reason—they have no operating picture of how work actually flows. You have spent fourteen years building exactly that picture into the Now Platform. The failure rate everyone else reads as a warning is, for you, the proof that the category you have been quietly building is real.

The Failure Rate Is the Proof, Not the Problem

MIT’s NANDA research documents that the large majority of generative-AI pilots fail to scale. The common thread is the absence of an operating picture of work—no map of which processes connect, which people own what, which systems depend on which. You have spent fourteen years building precisely that picture into the Now Platform. So the failure rate is not a problem to go solve. It is the evidence that what you have been building all along is a category, not a feature.

The Orchestration Layer, Not the Intelligence Layer

Every AI lab is racing to build the smartest model, converging on the same capabilities and racing one another toward zero on price. What is becoming scarce is the opposite thing—the operational infrastructure that makes intelligence reliable, governable, and accountable at enterprise scale. That is what your platform team builds. It means ServiceNow is not in the race the labs are running; it is the ground all of them need to stand on. The category name makes that legible to the market in a sentence.

When AI Stops Being an Add-On

ServiceNow raised its 2026 AI ACV target from $1 billion to $1.5 billion—your own number, your own ambition. But “AI ACV” is still a feature metric. The larger truth is the whole subscription base understood as infrastructure. When the market sees ServiceNow as the Work-Native platform rather than an AI-enhanced SaaS tool, the AI revenue stops being an add-on and becomes the proof of the platform itself. The number was never the point—it was always the evidence.