The most telling data point from ServiceNow's Knowledge 2026 conference in Las Vegas was not in any keynote slide or product demo. It was the gap between what the company was announcing and what its customers were asking for.
ServiceNow used the event to showcase its AI Control Tower and Now Assist platforms, positioning artificial intelligence as the centrepiece of its product strategy. Customers who attended responded with genuine interest, but their questions were not about the new features.
They wanted to know how to get more value from the products they had already bought, whether anyone could explain the licensing model clearly, and when they might see a measurable return on their existing AI investments.
The pattern is not unique to ServiceNow. It is an industry-wide problem in which enterprise software vendors are shipping AI features faster than their own sales teams can explain them, leaving customers to navigate overlapping products, unclear pricing structures and marketing language that substitutes enthusiasm for specificity.
Enterprise technology adviser Adam Mansfield, who works directly with large buyers, said customers are spending significant time trying to understand licensing models and costs but are struggling to get clear answers. Vendors, he said, are reluctant to offer volume discounts or commit to transparent pricing frameworks, creating friction in sales processes that were previously straightforward.
The shift towards consumption-based pricing is a particular source of anxiety. Under these models, customers pay based on how much they use a product rather than a fixed licence fee, which makes costs difficult to forecast and creates the perception, justified or not, that vendors are building a metered revenue stream on top of data that the customer already owns.
ServiceNow operates a hybrid model that blends traditional subscriptions with usage-based charges, but customers at the conference said they found the structure opaque and were worried about unpredictable cost increases as AI agents begin processing larger volumes of their data.
Despite the frustrations, no customers at the event said they were considering leaving ServiceNow. The company's core IT service management, customer service and HR platforms remain deeply embedded in enterprise workflows, and the switching costs are prohibitive. The stock, down roughly 39% year to date, reflects investor concern about the pace of AI monetisation, but the customer base itself appears stable.
Related reading
- ServiceNow posts Q4 and full-year 2025 results
- OpenAI's $18bn chip deal with Broadcom hits financing snag as Microsoft declines to commit
- ServiceNow integrates Claude AI across platform to drive enterprise automation
The message from the conference floor was not that AI is unwanted. It was that the gap between what vendors are promising and what customers can realistically evaluate, procure and deploy is growing rather than shrinking. Buyers want fewer new features and more clarity about the ones they already have: what they cost, what they do, and what measurable improvement they will deliver within a defined timeframe.
For the enterprise software sector, the lesson is uncomfortable. The companies that win the AI transition may not be those that announce the most products, but those that help their customers understand, with specificity and honesty, what they are buying and why it is worth the price. That is a harder sell than a keynote demo, but it is the one that will determine whether AI adoption accelerates or stalls.