The questions that matter: A leader's guide to early-stage AI transformation

kelliabbott · 3 weeks ago

Welcome back to our series from the AI Center of Excellence (CoE) team at ServiceNow. Through countless advisory and hands-on engagements, we've gathered valuable insights and practical guidance that we're excited to share with the broader ServiceNow community.

Here's what I've noticed in conversations with customers navigating their AI journeys: the leaders who succeed aren't necessarily the ones with the biggest budgets or the most technical expertise. They're the ones asking the right questions.

Let me show you what I mean.

Why Questions Matter More Than Answers

Here's an uncomfortable truth: according to New Vantage Partners, fewer than 20% of executives report having established a data culture—despite 97% investing in data initiatives. Think about that gap. Almost everyone is investing, but only one in five has built the foundation those investments need to succeed.

Many teams are building AI solutions on foundations that can't support them. Inadequate data infrastructure. Siloed teams. Misaligned incentives. The difference between organizations that thrive and those that struggle often comes down to whether leaders asked the right questions before writing the first line of code.

Here's the thing: AI transformation isn't primarily a technical challenge—it's a change management challenge. The technology is only as good as your team's ability to adopt it, trust it, and integrate it into their daily work.

That's why your role as a leader isn't to have all the answers. It's to ask the questions that surface the real barriers, align your team around shared goals, and create the conditions for success.

Frame the Change: Questions That Build Alignment

Before you evaluate tool selection or technical architecture, you need to shape the narrative around why this transformation matters.

I call this "framing," and it's about connecting AI initiatives to what your organization already values—whether that's customer experience, operational excellence, or innovation. When people understand why AI matters to them personally, resistance gives way to engagement.

Questions to ask your team:

"How does this AI initiative reinforce what we already do best?" This grounds AI in your organization's existing strengths rather than positioning it as a disruptive threat. For example, if your team prides itself on customer relationships, frame AI as "giving you more time for the conversations that matter" by handling routine inquiries.
"What would success look like for your role specifically?" AI affects different functions differently. Your service desk might celebrate deflection rates while your knowledge managers care about article quality. Understanding individual perspectives helps you address concerns before they become roadblocks.
"Where in your workflow do you spend time on tasks that feel like they should be easier?" This is my favorite question because it surfaces real pain points that AI can address. Recently, a customer answered: "I spend 30 minutes every morning manually categorizing tickets." That became their first use case—and their strongest advocate.
"What concerns do you have that we haven't addressed yet?" Creating psychological safety for honest feedback prevents underground resistance. You want concerns voiced in meetings, not whispered in hallways.

The goal of framing isn't to sell AI to your team—it's to co-create a vision where AI enhances rather than threatens their professional identity. People who feel ownership over the transformation become your strongest champions.

Structure for Success: Questions That Build Your Foundation

Once you've established the "why," you need to build the operational infrastructure to support execution. This is where good intentions meet reality.

Questions to ask your team:

"Do we have protected time for learning and experimentation, or will this compete with existing priorities?" Without dedicated time, teams default to old habits under pressure. I've seen this pattern repeatedly: great AI tools sitting unused because "I'm too busy." Build learning time into the schedule, or it won't happen.
"Who owns the outcome of this initiative, and is that clear to everyone?" Ambiguous accountability kills AI projects. Not "the AI team" or "IT"—which specific person is accountable for adoption? For business value? For ongoing improvement? Write it down.
"What cross-functional collaboration does this require, and have we built those bridges?" AI initiatives often require data scientists, domain experts, and engineers to work together in new ways. If these groups haven't collaborated before, you need to create the forums and norms that enable it.
"What barriers exist to data sharing across departments?" Data silos are one of the most common (and underestimated) obstacles to AI success. Sometimes they're technical. Often they're political. Always surface them early.

A note on organizational models: Organizations typically choose between two structural approaches for AI teams. In a service-oriented model, data teams report to functional departments—this works well for incremental improvements but can limit transformative potential. In an executive-led model, data teams report to the C-suite—this enables transformative innovation but requires strong executive sponsorship. Neither is universally "right"; the key is intentionally choosing the model that matches your ambitions.

Select the Right Projects: Questions That Maximize Early Wins

Not all AI projects are created equal, especially in the early stages. The projects you choose first will either build momentum and credibility or create skepticism that's hard to overcome.

Smart project selection is one of the highest-impact decisions you'll make.

Questions to ask before committing:

"Does this project address a real pain point that people actually care about?" Technical elegance means nothing if the solution doesn't solve a problem people feel. I've seen teams build sophisticated models for problems nobody had.
"Can we demonstrate value within 120 days?" Early-stage projects should build credibility quickly. Save the multi-year transformations for after you've proven the concept. Quick wins create the political capital for bigger bets.
"Do we have the data we need, and is it accessible?" The most common reason AI projects fail isn't the algorithm—it's the data. Before you commit, actually look at the data. Not what you think you have. What you actually have.
"What's the risk if the AI makes a mistake?" Start with use cases where errors are recoverable. High-stakes decisions should wait until you've built both capability and trust. Password resets? Low risk. Loan approvals? Save that for later.
"Will this project create advocates or skeptics?" Choose projects that touch people who will champion success—their enthusiasm is contagious. One excited early adopter is worth ten skeptics you convinced.

Set Meaningful KPIs: Questions That Measure Real Progress

Measuring AI success requires new leadership skills. Traditional metrics often miss what matters most in early-stage transformation: learning, adoption, and capability building.

Questions to guide your measurement strategy:

"What would convince a skeptic that this is working?" This forces you to identify evidence that matters to your toughest critics. Not what impresses you—what would move them?
"Are we measuring adoption or just availability?" A tool that exists but isn't used delivers zero value. Track usage, not just deployment. How many people logged in? How many used it this week? How many found it valuable enough to use again?
"What are we learning, and how are we capturing those lessons?" In early stages, speed to learning matters more than speed to delivery. Document what worked, what didn't, and why. This becomes institutional memory.
"Can we articulate the ROI in terms executives care about?" Time savings, error reduction, and customer satisfaction translate better than technical metrics. "Reduced ticket resolution time by 23%" beats "achieved 94% model accuracy."
"What trade-offs are we making, and do all stakeholders understand them?" Every AI decision involves trade-offs between accuracy, speed, cost, and risk. Transparent discussion prevents later surprises. Better to discuss trade-offs upfront than defend them after launch.

Earn Buy-In: Questions That Address the Human Side

Here's what I've learned: resistance to AI often isn't about the technology—it's about fear. Fear of job loss. Fear of losing control. Fear of being made irrelevant.

The most effective AI leaders address these human concerns directly rather than dismissing them.

Questions to build trust:

"How can we position this as augmentation rather than replacement?" The most successful AI implementations enhance human capabilities rather than automating humans away. Frame it as "handling the work nobody wants so you can focus on what matters."
"What new skills will our team need, and how will we help them develop those skills?" Investing in reskilling demonstrates commitment to your people's futures. Don't just say "don't worry." Show them the plan. Show them the investment. Show them the path forward.
"How are we involving end users in the development process?" People support what they help create. Early involvement transforms potential resisters into co-designers. Some of your best insights will come from end users, not technical teams.
"Where should humans maintain decision authority?" Clear boundaries about human oversight build trust and ensure accountability. "AI recommends, humans decide" for high-stakes scenarios. Say it clearly. Mean it genuinely.
"What feedback mechanisms exist for people to raise concerns or suggest improvements?" Continuous feedback loops signal that you value human judgment alongside AI capabilities. Make it easy to flag issues. Make it safe to admit concerns. Make it visible that you act on feedback.

Putting It All Together

Early-stage AI transformation succeeds when leaders focus less on having answers and more on asking questions that surface truth, build alignment, and create conditions for learning.

The questions in this article aren't exhaustive—they're starting points designed to spark the conversations that matter.

Key takeaways:

Frame AI as enhancement, not threat. Connect initiatives to existing values and help people see themselves in the transformation.
Structure deliberately. Protected time, clear accountability, and cross-functional collaboration don't happen by accident.
Choose projects strategically. Early wins build momentum; early failures create skepticism that's hard to overcome.
Measure what matters. Adoption, learning, and capability building matter more than technical metrics in early stages.
Address fear directly. Human concerns deserve human responses—not dismissal.
Develop yourself. Leadership at any level means continuously growing your own capabilities.

Your Next Step

The organizations that succeed with AI transformation won't be the ones with the most sophisticated algorithms or the largest data sets. They'll be the ones led by people—at every level—who ask the right questions, listen to the answers, and create environments where learning and experimentation can flourish.

Here's what to do this week: Choose one section from this article that resonates most with your current challenges. Schedule 30 minutes with your team to discuss just those questions. Document what you learn.

Small, consistent conversations create the foundation for successful transformation.

If you have questions or thoughts, drop them in the comments—we'll respond or update the article as needed. If you found this helpful, please share your feedback or link to it on your preferred platform. This is just one part of our series on AI—stay tuned for more.

For tailored guidance, reach out to your ServiceNow account team.

Views expressed are my own and do not represent ServiceNow, my team, partners, or customers.

https://www.businesswire.com/news/home/20220103005036/en/NewVantage-Partners-Releases-2022-Data-And-...