Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.
Back to Podcast Digest
How Agentic AI Changes The Workplace — With Jacqui Canney and Kellie Romack
TL;DR
ServiceNow doubled HR’s capacity without layoffs — Jacqui Canney said HR business partners went from supporting roughly 400 employees each to 1,000 by using AI on parts of jobs, then redeploying staff into higher-order work, AI training, and new use cases as the company grew from about 14,000 to nearly 30,000 employees.
The best AI wins came from redesigning workflows, not stapling AI onto old ones — Kellie Romack’s standout example was a commissions process that took 4 days and now takes 8 seconds, because ServiceNow reimagined the whole cross-functional workflow across finance, HR, and sales instead of just speeding up one step.
Agentic AI is already handling most routine internal support work at scale — Romack said ServiceNow’s IT service desk now resolves 90% of requests autonomously from first touch to resolution, and 85% of the people who used to do that work were moved into roles like SecOps, AIOps, and executive briefing centers.
Managers are becoming both people leaders and agent managers — the conversation pushed back on the idea that AI kills management entirely, arguing instead that humans still decide what should be built, what gets governed, and how teams handle complex edge cases while supervising an “agentic workforce.”
Training only works if the company gives people permission to use AI at work — Canney described “AI required work days,” manager enablement, role-based AI skills assessments, and ServiceNow University, which they say has 1.88 million learners on the way to 3 million.
Governance is becoming a core product and operating requirement — ServiceNow built an internal “AI control tower” to monitor agents, adoption, value, and security across the company, then turned it into a product after realizing how easily duplicate agents, shadow development, and token costs can spiral without visibility.
The Breakdown
From theory to production: what AI at work actually looks like
Alex Kantrowitz opens by steering the conversation away from speculation and toward what’s already live. ServiceNow executives Jacqui Canney and Kellie Romack make it clear right away: this isn’t a future-of-work panel, it’s a “we have things in production today” conversation.
HR asked employees for ideas — and got 1,000 of them
Canney says one of the first smart moves was simply asking employees where AI could help, which generated 1,000 use-case ideas fast. From there, HR prioritized, deployed agents into people operations, and more than doubled output without adding headcount — but notably, they didn’t cut the people doing the work; they moved them into tougher cases, AI training, and building more automation.
The commissions story: 4 days collapsed into 8 seconds
Romack gives the most concrete example in the video: employees and sellers were flooding finance with commission-status tickets because people understandably care about how they get paid. Instead of patching the old process, ServiceNow rebuilt it end-to-end with AI and security, cutting turnaround from four days to eight seconds — a huge quality-of-life win for finance and sales alike.
AI “disintegrates the org chart” by attacking the problem, not the silo
Canney’s line here is memorable: AI is “disintegrating the org chart.” Her point is that the old model solved for HR problems, finance problems, or operations problems separately; the new model starts with the employee problem itself, then crosses functions to fix it.
Upskilling only works if managers and culture make room for it
The conversation shifts from tools to permission. Canney says ServiceNow has “AI required work days,” invests heavily in manager capability, and creates space for experimentation so employees don’t feel punished for trying; Romack adds that this pilot-heavy approach is how they’ve reached 95% AI adoption across the company.
What happens to managers when agents take over routine work?
Asked whether middle management could disappear, Romack says it’s more complicated than that and offers her own IT service desk as proof. ServiceNow automated 90% of service desk work end-to-end, then moved 85% of staff into higher-value roles, while the remaining team now handles complex cases and manages the agentic workforce — so the job changes, but human coordination, judgment, and escalation don’t vanish.
Careers get wider, not just higher
Canney pushes back on the assumption that everyone wants a neat, linear climb up the org chart, arguing that people increasingly want different things and employers need to be transparent about what kind of company they’re building. Romack echoes that in practical terms: redeployment can mean moving into security or AIOps, and “growing your career means getting wider,” not just getting promoted.
Governance, content, and the control tower behind the scenes
One of the most useful operational details is ServiceNow’s internal AI control tower, built first out of necessity to monitor agents, value, adoption, governance, and security across the business. They also stress that success depends on high-quality content, human-in-the-loop monitoring, and safe experimentation — enough that an HR-built benchmarking agent inside a governed environment became so valuable it’s now being turned into a customer-facing product.
Share
Keep Reading
Two Ways to Follow
The Weekly Echo. The inbox-shaped summary of what mattered.
New editorials announced here.
