ServiceNow’s John Phillips: Stop Measuring AI Adoption, Measure Work Outcomes

As organizations grapple with the rapid proliferation of AI tools, ServiceNow’s Group Vice President of Employee Experience, John Phillips, is urging a fundamental shift in how success is measured. In a recent episode of the podcast You Should Know, Phillips argued that counting how often employees use AI tools is a flawed metric. Instead, he says, the focus should be on whether jobs are completed faster, with less friction, and with better outcomes for both employees and the business.

The conversation comes at a time when chief human resources officers (CHROs) face mounting pressure to demonstrate AI’s productivity gains across increasingly fragmented technology stacks. Phillips described the current landscape as a “train wreck of productivity,” noting that “every system of record is now got their little AI agent and it's creating chaos for these practitioners.” He predicts a rapid shift in measurement: “We're going to quickly stop talking about AI adoption as tool usage, and we're going to start looking at the outcomes and jobs to be done.”

Phillips’ remarks highlight a growing concern that AI tools, rather than streamlining work, are adding complexity. Many organizations have purchased multiple AI solutions that do not integrate, with customers often arriving with “eight purchased AI tools plus one they built themselves, none of which talk to each other.” This fragmentation undermines the potential for AI to deliver on its promise of enhanced productivity.

The episode also explored the two-sided value exchange between employee and employer, and what happens to the 23 hours a tool claims to save. Phillips emphasized the importance of discretionary effort over traditional engagement surveys, suggesting that hyper-personalization is more effective than one-size-fits-all pulse data. He also touched on the collapse of work boundaries post-COVID, which has contributed to burnout and an internal dialogue of “am I enough.”

ServiceNow’s approach involves an AI control tower vision—an agentic overlay that stitches together 15 large language models (LLMs) and 100 systems, rather than ripping and replacing existing infrastructure. This strategy aims to create a unified experience for employees, reducing the chaos caused by disparate AI agents.

Phillips also connected his philosophy to his experiences in refugee camps, arguing that “skills and talent is universal and opportunity is not.” This perspective underscores the importance of designing AI and work processes that are inclusive and equitable.

The discussion resonated with co-hosts Ryan Leary and William Tincup, who have long critiqued engagement surveys. Tincup pushed Phillips on whether discretionary effort is the truer metric, while Leary shared a personal anecdote about applying to Home Depot and never receiving an acknowledgment email, illustrating the need for better employee experiences.

As AI continues to permeate the workplace, Phillips’ call to focus on outcomes rather than adoption is a timely reminder that technology should serve human performance, not the other way around. High performance, he noted, “has both extreme focus and extreme recovery. It's in any environment, the highest performers in the world have those things.”