ServiceNow CDIO's 5 steps to AI success

Kellie Romack, chief digital information officer at workforce automation company ServiceNow, is a well of advice for CIOs rolling out AI initiatives, starting with her biggest prohibition -- automating legacy processes is not the right starting point. 

"The advice that I give other CIOs is you have to understand the user, understand the outcomes, understand the value -- don't just start with the myopic," Romack said, speaking at Gartner's recent IT Symposium/Xpo in Orlando, Fla. 

Romack joined ServiceNow in 2022 with a focus of "streamlining internal operations," and took on the role of CDIO about 18 months ago. As CDIO, she has focused on delivering enterprise-wide digital transformation and improving both the employee and customer experiences with AI.

In addition to advising CIOs to resist the temptation to just automate old processes with AI, Romack also recommends gaining buy-in from the C-suite, focusing on business outcomes, unifying the AI experience and making AI education mandatory for a successful AI deployment. 

Here is her annotated to-do list. 

1. Reimagine, don't just automate, work 

While it can be tempting to layer AI over existing processes, that isn't the correct approach, Romack said. With AI implementations, "we have to reimagine the work," she said. "You want to try to innovate to the new," not automate the old. 

Related:The AI orchestration gap: What business leaders must fix

She encourages CIOs to instead "zoom out" and take a hard look at what the business value will be for any AI deployment. That's the approach she takes when one department might propose a use case for AI, but it unfortunately won't scale to benefit the rest of the organization. 

"I have those difficult conversations internally, because I'm responsible for harmonizing everything," Romack said. 

(Source: ServiceNow) Kellie Romack, CDIO, ServiceNow

2. Achieve top-down buy-in

CIOs deploying AI in their organization cannot proceed without buy-in from the rest of the C-suite, Romack said. Once a month, she has a meeting with all the C-suite executives at ServiceNow. Dubbed the "C-Suite Inspection Meeting," its purpose is to discuss AI adoption, where ServiceNow is on its AI journey and how the executives are using AI. 

The executives need to understand the gains from AI, she said, but the meeting is not just about value. "Value, scale and outcomes are very important, but it's how they're using [AI] and what they're using it for," she said.

ServiceNow also has a company-wide AI council of about 20 leaders representing every department, Romack said. 

"Everybody learns from each other, too. We're prioritizing different areas so that we can make sure we get the most scale," she said. 

Related:Model Context Protocol servers: Build or buy?

By prioritizing scale, Romack means that she must weigh the AI needs presented by the council and determine which use cases will have the greatest impact for the most employees.

"Especially things that are repeatable -- of course, we should use AI to do that," she said. "Then we could personalize [services], because AI is so good at that." Creating a repeatable framework and tailoring it to individual needs applies to customers, partners and users. 

3. Focus on outcomes

Romack said focusing on AI outcomes not only pertains to the business value the company as a whole is trying to achieve or addressing the needs of the majority of employees, but also understanding the specific problems and desired outcomes of her business peers in the C-suite. 

"They're coming to me with a business problem, and it's our job to bring those solutions," Romack said. 

4. Create a unified experience

Achieving AI adoption and scale requires creating a seamless and cohesive user experience, she said.

"If you don't have a great experience that unifies it all, you're not going to get the adoption, which will not give you the scale," she said.

ServiceNow launched "AI Control Tower" about a year ago as a centralized hub and dashboard for customers to manage their AI technologies across their organization. On the dashboard, customers can view their AI strategy, AI asset inventory, track the value of their AI, monitor security and more in one location. 

Related:City of Raleigh CIO's 'Crawl, Walk, Run' Approach to AI

AI Control Tower also tracks the performance of internal AI use cases. Within AI Control Tower, ServiceNow launched AI Agents for CRM Customer Support. ServiceNow developed the AI Agents by meeting with service agents and support engineers. The ServiceNow team also listened in on calls to better understand customers' requirements for managing AI -- and started with the company's most complex IT service tickets. Using AI Agents for CRM Customer Support, the company was able to reduce its response time by 50%, Romack said.

5. Make AI education mandatory

AI education has been made available to the whole organization -- and it is mandatory for everyone, she said. 

"I'm a big educator, so we moved AI from a 'black box' to a 'glass box,' and then helped them understand what we're using, how we get the value, and how we're scaling it for them, and with them," Romack said. 

The organization also provides AI certification opportunities according to employee roles and requirements. ServiceNow University is another continuing education resource available to customers and partners, offering some free content. According to Romack, the company aims to upskill 3 million people at ServiceNow University during the next two years.