Shelve Your Ego, Own Your Output: Lessons From Leaders Putting AI to Work

Enterprise leaders share their top AI concerns, where agents are paying off, and how to get started.

Two years ago, most enterprise AI conversations centered on access: who got a license, which model to use, how many tokens employees could burn though, and which pilot to fund. But now the questions are different. How do you scale AI productivity across teams? How do you keep people accountable for work an agent helped produce? And how do you bring an entire workforce through such rapid change?

At Workday Rising’s Women in AI event last month, I met leaders, spanning technology, HR, finance, and marketing, who discussed the challenges they face implementing AI capabilities across their organizations. Despite coming from different industries, regions, company sizes, and AI maturity levels, they found consistent starting points, hurdles, and strategies. 

In a packed convention center, executives agreed AI adoption starts small and personal, and judgement and accountability are the biggest concerns. From first experiments to company-wide software migrations, here’s what they said worked, what needs iterating, and practical starting points for leaders in the early stages.

Where Teams Are Starting

Leaders agreed the easiest place to deploy enterprise AI is in a simple daily task. Many started by asking AI to draft emails and studying the results. Because AI reflects you to you, this can reveal habits you didn’t know you had, like over-used phrases or misconstrued voice or tone. It gives you the chance to modify yourself as you teach AI to do tasks for you. These initial steps are important before handing AI something more consequential. It teaches you not only how you come across, but also how to prompt AI to get the results you want. 

Employees learning on their own is a great start, but scaling that testing and learning across teams is the next step. Leaders have learned not to expect all employees  to volunteer questions. So adding some structure to shared learning provides a space to try new things and see others’ results. Task forces to try out use cases and hackathons to come up with ideas and solutions can make the process more collaborative. Even rotating a slot in a weekly meeting where a team member shares how they used AI that week can make adoption faster and more fun together. 

Global financial powerhouse FIS built AI adoption around a mindset it calls “Shelve Your Ego.” FIS senior leaders are encouraged to adopt a beginner’s mindset and ask colleagues questions they might not have asked before. This comes to life through a mentoring program matching early-career, AI-native employees with senior executives. The junior employees coach leaders on the technology, while executives share judgment, critical thinking, and an understanding of how the organization works. 

The Concerns Leaders Are Weighing

No matter the industry, every leader responsible for a team expressed concern about the responsible use of AI, tracking it, and assessing its impact. The biggest concerns fell into a few areas:

  • Accountability. The most common worry is people outsourcing their thinking. One HR leader said when her team first got access to an LLM, she received so many documents full of vague buzzwords that her company updated its policy to say: employees own anything AI helps them produce. Creating AI policies that make human authors accountable for their AI generated content makes the content more human and the responsibility clear.

Employees own anything AI helps them produce.

  • The pull toward doing less. Once a tool takes over part of a task, people tend to disengage and get distracted. If AI does the easy work, then the work left for people is harder.

  • An AI that mirrors you. AI tends to validate your opinions and ideas. If you don’t prompt it to point out what is missing in your instructions, you risk bias as well as results that read like unedited AI slop.

  • Traceability. AI generated content is only as good as the data and material it can access, so making sure yours is current and clean gives you more usable results. One leader asked AI to summarize figures from 65 presentation decks and got numbers it couldn’t trace to any source.

  • Sprawl and governance. Every platform now ships its own AI, and agents multiply fast. Leaders want a single front door, a central hub to track agents, and visibility into token usage and consumption.

  • Making the business case. It takes more than a story to sell an investment to a leadership team. AI is new enough that while people understand the value, pitching solid metrics to get more tools is still hard to do.

The good news: leaders who’ve worked through these concerns have results to show.

Where Agents Are Delivering

Workday’s IT and HR teams have tested agents to tackle our own business challenges. Ashley McCord, Workday’s senior vice president of enterprise technology services, prompted an agent in three sentences to work through her ServiceNow and Jira queues. The agent closed 5,000 tickets in four days and surfaced the root cause her team had missed: tickets routed to the wrong queues and people.

Ashley used the same approach when Workday moved 20,000 workmates from Outlook to Gmail and her small team faced a weekend flood of questions. She built an agent that found unanswered questions in the Slack support channel and replied in her voice, grounded in the team’s FAQs and expert answers. The agent answered more than 2,000 questions that weekend and saved the team 80% of their time over the following weeks.

“If I can do this as a leader, think about what my team could do.” 

Ashley McCord SVP, Enterprise Technology Services Workday

Before that migration, Workday CIO Rani Johnson had to win over a leadership team worried about change fatigue. She paired a survey of nearly 7,000 Workday employees with Sana’s analysis of Slack, email, and direct messages. She then used the results to plan a phased rollout. Sana kept listening and making recommendations, so the team could adjust before each wave.

Outside IT, Workday’s HR team built a check-in agent to draft quarterly self-appraisals and manager reviews from a quarter’s worth of Slack, email, and documents. The agent delivered 80% time-return and a 50% improvement in review quality, and it’s now Workday’s most-used agent.

In marketing, the Workday customer advocacy team built a customer story agent that applies brand messaging, fills the digital team’s template, and adds SEO and AEO tags, cutting the work by more than half.

AI can also be a thought partner that challenges presumptions, organizes information, and finds gaps in strategy. One leader asked it to take on personas, such as a CFO, CIO, CEO, or frontline worker, to see how a message would land with each audience.

What these agents all have in common is a well-defined set of instructions, a safe testing ground, and parameters on what they are designed to do. You can’t and won’t be an expert on day one. The leaders who got furthest were willing to start as beginners, share what they learned, and iterate with their teams. If you’re just starting out, here’s where they suggest beginning.

Starting Points for Success

If you’re just getting started piloting and adopting AI capabilities, here a few tips and tricks: 

  • Start personal, then go after your hardest task. Low-stakes experiments build confidence. The value grows when you aim agents at the work that costs you the most time.

  • Put building in the hands of the people closest to the work. Process owners know where the friction is. Start off with agents focused on one business problem instead of trying to automate an entire function.

  • Track agents in one place. A central hub prevents duplicate builds and makes good agents easy to share.

  • Write accountability into policy. Make it clear that the person who sends the work owns it, however it was compiled.

  • Measure results to earn the next investment. Hours saved, tickets closed, and quality gains turn an exciting story into a business case.

  • Mentor in both directions. Pair AI-fluent early-career employees with experienced leaders, so fluency and judgment travel both ways.

  • Redesign teams with data. Start from the outcome you want to change, base it on clean data, and make intentional choices about how teams and work should evolve.

The leaders seeing results treat AI less as a tool and more as a teammate, one that frees time for the work they value most: leading, deciding, mentoring, and building relationships. The opportunity is yours to experiment, pilot, share what you learn, and keep human judgment at the center.

For more details on the skills and mindset that make AI easier and faster to adopt, keep reading.

Check out What’s a Polymath and Why are They on the Rise?

More Reading