$NOW processes 85 million AI-assisted workflows per day. That's not a chatbot demo. That's infrastructure.
While the market was debating which LLM wins the foundation model race, ServiceNow quietly embedded AI into every IT ticket, HR request, procurement approval, and compliance workflow at 8,000+ enterprise customers. The distribution advantage they've built is something OpenAI and Anthropic can't replicate with a better model.
───
Here's the frame I use for enterprise software right now: there are two types of companies in the AI transition.
Type 1: Companies that sell AI as a feature — a copilot add-on, a smart search bolt-on, a summary tool you can toggle off.
Type 2: Companies where AI is the workflow itself. You can't separate the AI from the product because the AI is the product.
ServiceNow is a Type 2 company. And the market is still partly pricing it like a Type 1.
───
The bull case here isn't complicated, but it's easy to underestimate.
Enterprise software has always been about one thing: reducing friction in how large organizations operate. ServiceNow has owned IT service management for 15 years. Now they're expanding that footprint into HR, finance, legal, and customer operations — using AI agents to execute workflows, not just suggest them.
The difference between "suggest" and "execute" is enormous.
When AI suggests an action, you still need a human to approve, route, and close the ticket. The headcount savings are marginal. When AI executes — creates the incident, routes it, resolves it, closes it, updates the audit log — you've just eliminated 3–5 touches on every workflow.
ServiceNow's Now Assist product is doing this at scale. Their customer data shows 40–60% reduction in ticket resolution time at major deployments. That's not a soft productivity metric — that's headcount avoidance showing up in CFO budget models.
───
Here's why I'm watching this specifically right now.
Enterprise software buying cycles are long. The deals signed in 2024 and 2025 — when CIOs were cautiously piloting AI — are starting to expand into full platform rollouts. We're entering the "land and expand" phase.
ServiceNow's net expansion rate has historically run at 125%+. In an AI execution cycle, that number can move higher because the value realization is faster and more quantifiable. When a CIO can show their CFO a dashboard with "tickets per agent per day: +68% since Now Assist deployment," budget approval for the next expansion is easy.
That's different from the old enterprise software pitch, which was always "trust us, productivity will improve in 12–18 months." Now it's measurable, immediate, and directly tied to headcount costs that CFOs are desperately trying to control.
───
The valuation question is fair. NOW trades at 15–17x forward revenue. That's not cheap by any standard.
But here's the math that matters: if ServiceNow can sustain 20%+ revenue growth for the next four years (their current trajectory), and improve free cash flow margins from ~28% today toward 35%+, the stock doesn't need multiple expansion to work. It just needs the business to execute.
For an enterprise software company at their scale — $12B+ in annual revenue — growing 20% is not trivial. It requires real share gains and real product differentiation. The AI execution layer is what gives them the right to grow into those numbers.
───
The risk I watch most: Microsoft.
Teams, Copilot, and the Microsoft 365 stack are deeply embedded in the same enterprises ServiceNow serves. If Microsoft builds a credible workflow execution layer on top of their productivity suite, ServiceNow's expansion thesis slows. That's the bear case worth taking seriously.
So far, Microsoft's AI workflow story is still mostly in copilot-suggests territory. ServiceNow is ahead on execution. But the gap could close.
NOW is a name I think about differently than most enterprise software. It's not just a SaaS company riding the AI wave.
English
