Ask a room of Platform Owners what's standing between their organization and AI, and you'll hear about models, licenses, and roadmaps. Ask three who are doing the work, and the answer changes. At our first Platform Owner Academy, more than 100 of you joined 3 Platform Owners -- Derek Nohr (Elara Caring, formerly Avalara), Joy Coats (Methodist Le Bonheur Healthcare), and Katie Green (St. Jude Children's Research Hospital) to talk about getting your platform AI-ready. The conversation kept coming back to the same idea: AI readiness depends on trust. That means trust from stakeholders, trust in your data, and trust in the numbers you take to your executives.
Trust starts with one visible win
Katie's first lesson at St. Jude was about a spreadsheet. The disaster recovery team was tracking tier and business criticality outside the platform. Her team moved that information into the CMDB and made sure it was populated whenever a new record was created. Nobody argues about how critical a system is anymore, because the answer is in one place, and the data is governed where it enters. "CMDB is boring," she said, "but if you can have good information in there that's helpful, it's great."
Joy found the same thing with knowledge management. She didn't lead with the platform's benefits. She showed knowledge owners their own payoff: fewer incidents landing on their desks and more say in what the article says. A few proof-of-concept areas got it working, and other teams asked to join.
Derek described what happens next. After a few hesitant stakeholders saw results, he started getting calls from people he'd never met, including someone with a sprawling vendor risk spreadsheet who wanted to know what could be done. Win one stakeholder, make the result visible, and the story travels. He also offered a warning: trust is hard to build and easy to lose, and a team seen as a black hole ends up pushing a rock up a mountain.
A webinar attendee asked the question many POs face: stakeholders have day jobs, so how do you get them on board? Two answers stood out. Derek and Katie both bring in peers from ServiceNow's Now on Now program, people in finance, procurement, or legal who use the platform daily. The pitch comes from someone who knows the stakeholder's pain, not from the Platform Owner. Katie brought ServiceNow's legal team in to talk with St. Jude's legal team because "they speak a different language." Joy added that much resistance is plain fear of change, so understand the dilemma before you sell the value. When another attendee asked how to grow adoption when most people don't know what ITSM is, the answers were the same: real stories from peers who've solved it, and executive sponsorship so the message is consistent from the top down.
Your data will tell you the truth about your AI
Trust in people is one thing. Trust in data is another, and it was the most energetic part of the session.
Derek learned his lesson the hard way. When he deployed Now Assist AI agents, they were "kind of worthless" because the data behind them was thin. Assignment groups were stale, and some new applications had none at all. He only found the missing fields once the agents "started acting crazy." His advice is not to wait for clean data: you'll never reach 100%, and the gaps are easier to find once the AI is running, so fail fast.
Katie takes a deliberate approach. She launched an effort she calls “Rebaseline”. It resets the fundamentals, like your principal classes and what "support group" actually means. It names data owners who help identify sources and important fields, and it tunes dashboards to show progress from X% to Y%. Her rule is that this is not a project. "It will never be done." She's focused on completeness, correctness, and compliance, and she's hunting for swag budget to run a cleanup competition between teams, because the work is boring and boring work needs a little fun. Joy's team is taking a different route. They're in an assess phase, using the Impact team and accelerators to educate themselves on what they want before they draw up a plan.
Then came the question that hovered over the whole hour: should you start over? Is anyone re-platforming to get back to out-of-the-box?
One webinar attendee went off mute and said he has worked through Impact assessments, and his CMDB scored well on paper. In practice, he says, his data is atrocious. His team is split 50/50 between deleting the CMDB and fixing forward. He favors fixing forward and suspects few organizations truly need a full re-platform, though starting over is often easier to rally people around. He also made a point that stuck with the room: calling it a "database" undersells it. "If you were to call it the heart of ServiceNow, it would probably get a little bit more affection."
Another webinar attendee, on the platform since 2012, did start fresh. After years of customization that eventually worked against them, his team kept only the user records. A clean slate let them draw a line, so requests like "can you turn the screen purple?" got a simple no. The effort took about six months including assessments, and the old instance sat read-only for roughly a year before it was retired. "Well worth it."
The two views share a point: re-platform or repair, you have to trust your foundation before you can build on it. On how to find out where you stand, the panel agreed. Derek once paid a partner $55,000 for a health scan and says Impact's health dashboards have made that unnecessary. Katie and Joy both rely on Impact scans and accelerators. Katie's 10-year-old instance raised the same "start over?" question, and the scans helped her choose to re-baseline first.
Proving value means speaking two languages
Derek calls this the defining challenge of the role, and it's what keeps him up at night: the executive asking why ServiceNow costs so much. His answer is that Platform Owners have to think two ways at once. The tactical side covers throughput, incidents closed, and resolution times. The business side covers deflection, cost avoided, headcount, profitability, and EBITDA. They're two different skill sets that have to live in one person. "If you're not able to translate that into business outcomes," he said, "there's a whole bunch of other tools that are a little bit cheaper that they're gonna start looking at."
Katie reinforced this. Working with her Impact squad and CSE, she chose a handful of value-based metrics leaders connect with, such as accelerating request fulfillment and optimizing IT and HR service delivery. "It can't just be, there's a thousand incidents closed in the last three months. That's not helpful." Joy's team goes beyond MTTR and SLAs. She used Impact analysis to show the security value of GRC, IRM, and BCM, and her asset management dashboard shows what devices exist and what can be recovered, so a new team can reuse equipment instead of ordering more.
Derek's most practical tip: bring your boss, and your boss's boss, into the exercise. He built his value model with his Impact CSM and his VP of IT. Together they settled on a blended labor rate and the KPIs she cared about. Because she helped build the numbers, she trusted them, and she was comfortable taking the deflection savings to the CEO.
Start before you feel ready
When asked for parting advice, each panelist said something similar. Derek: "Have a bias towards action. Just try it. You'll learn more from something that doesn't work than from waiting.” Joy: "Don't wait for perfect." Katie: "Lean on others that have already been there, and then also provide your way. It's not a one-way street."
That last line is why we built this community. The best answers don’t come from a slide deck. They come from peers who had already run into the problem.
Join us for the next Academy on October 28. ServiceNow University will lead a discussion with Platform Owner panelists on building internal ServiceNow expertise from the ground up. Sign up here: https://info.servicenow.com/ServiceNowExchange-PlatformOwnerAcademy.html
Help us answer these
We ran out of time before we could get to every question. These are the ones still open, and we'd love to hear how you've handled them. Share your experience in the comment!
- CMDB ownership: How do you build enthusiasm for assigning ownership when owners aren't easy to identify?
- AI adoption with governance: Where are the best opportunities to use AI while keeping governance intact?
- The business case for new licensing: How do you show your CIO or CFO enough value to justify moving to the new license model?
- Governance after re-platforming: How do you assess and reset the operating model so you don't end up back where you started?
- Shadow app building: Business teams are spinning up apps with outside AI and low-code tools. How are you positioning the platform in response?
- End-to-end agents: How would you approach building queues or agents that handle tickets from start to finish?
- Build Agent vs. Now Assist: How do the two compare on customization and governance?
- Agentic AI for ticket data: Which capabilities in upcoming releases can teams build with today, and which need new licenses?
- DevOps toolchain data: How do you make ServiceNow the system of record for change when engineering teams use tools like ADO and Git in very different ways?
