The AI Paradox and the Hidden Costs of Downtime

By Stefanie Hoffman, Content Marketing Manager

The challenge with AI is no longer access to the technology and has evolved to how to turn experimentation into measurable results. AI Leadership Institute founder Noelle Russell shares how organizations can scale AI responsibly while staying aligned to business goals.

With leadership experience spanning Amazon, Microsoft, and Accenture, she has spent more than a decade helping organizations navigate the transition from AI enthusiasm to responsible, scalable enterprise integration. Today, she works with executives around the world to build the skills, governance models, and operations practices needed to deploy AI successfully.

Leadership and Lessons is a new Perspectives interview series that explores the experiences, decisions, and insights that shape today’s technology leaders. In our debut installment, Perspectives sat down with Russell to discuss early-stage AI governance, ways organizations can move past an “AI vending machine” mindset into something more meaningful, and how to build AI agents that turn AI potential into a measurable competitive advantage.

Stefanie Hoffman: How did you get your start as an AI entrepreneur, technologist, and educator? What is the most significant “aha” moment in your career that led you to found the AI Leadership Institute?

Noelle Russell: My journey started as a principal cloud architect at Amazon. At the time, my son, who has Down syndrome, was nine years old and just starting to realize that he didn't talk like everyone else. I was looking for ways to just make his life better. It turned out that it was a great opportunity for me to jump on board with a product that I actually thought could help him.

As soon as I joined the Alexa team, I realized how very different I was from everyone else. I was looking through the lens of being a mom and a caregiver and everyone else was a good ten years younger than me. They just were in a different span of life. As a result, I wanted to build different things for the company. I ended up building over a hundred apps for Amazon Alexa in the first year.

The reason that I became successful is that I became a builder. Prior to that, I was a cloud architect, but I was not someone who would get an idea and go build a mobile app or an iPhone app. That wasn't my identity. Alexa really brought that out of me. If all I needed to do was use my words to say something and a model would do what I say, what would I build?

And I went crazy. I built healthcare apps, childcare apps, mom apps, meditation apps, prayer apps. I just got very excited at the idea.

Hoffman: How did that moment shape your vision for the future of the industry?

Russell: About halfway through my tenure at Alexa, I realized I was in this magical unicorn role where we had one foot in technology and one foot in the business. We were the translators. We got the business excited about what is possible in technology.

Talking to organizations, I realized that they had no idea how to use what we were building, or how to leverage it, or how it would create value.

So I started the AI Leadership Institute in 2015. It came out of the same need that every executive had. Many of them were just stuck in problems with data connectivity, data privacy, enterprise architecture, and cloud infrastructure. There were so many challenges that they couldn’t see the opportunities.

The AI leadership Institute was all about teaching the opportunity side of this technology. We get excited when working with this technology, and I wanted that to become the ethos of what AI would become for them.

Hoffman: What does the AI Leadership Institute offer that isn't provided anywhere else? What is the unique value proposition?

Russell: In my book, Scaling Responsible AI, I use the analogy of a baby tiger. Most of us have little baby AI systems all over our organizations.

What I found is that everyone gets really excited about AI very quickly.

When the AI system is really small, that's the best time to ask the most important questions. Baby tiger, how big are you going to get? What are you going to eat? How much are you going to eat? Are you going to pay for it?

But no one's asking those questions.

What is different about my approach is that I say you need to enjoy your enthusiasm. Every model starts as an enthusiastic little baby. But how do you encourage and provide psychological safety for your organization to ask really hard questions in the midst of the enthusiasm stage?

Then you have to transition to the practical, the responsible, the secure, and the private. The organizations that struggle most are often the ones that never make this transition. They just keep treating it like it's a cute little AI system, even though it's beginning to grow up and starting to do things that are out of their control.

Hoffman: Why is AI literacy important for everyone? How is the fundamental knowledge of AI going to transform communities? And in what ways?

Russell: AI Leadership Institute started as an executive education organization, but then during Covid, I developed a framework that helped people realize how to get value from their AI. It’s called ARC, like a leadership or maturity arc.

This ARC starts with Awareness. The first thing you have to do is just be aware of what you have, the consequences, the benefits, and the outcomes.

That awareness can't just be for technologists. In most cases, AI today is being driven by strategists, business owners, line-of-business leaders, and executives. That awareness of what AI is, what it does, how to do it safely, and how to be responsible has to permeate through the entire organization.

“R” stands for Readiness. In order for a company to actually develop readiness, first, everyone has to know what they're working on. Second, you have to put your fingers on the keyboard and build something. Learning AI on YouTube is fine, but you won’t really understand it until you build it. If you actually learn it, your whole organization becomes aware. You develop hands-on readiness through communities of practice.

The inevitable outcome is Competitive Advantage, which is the “C,” for my business, users, clients, and employees.

That small shift into personalization of AI is where all the value is.

Hoffman; What is the most common myth about AI literacy that you find yourself debunking?

Russell: One of the biggest myths I end up working through with executives and boards is this idea that AI is not inherently secure or safe.

Security is really about understanding your identity. What is your security boundary? What is secure to you?

Another big myth is that many organizations are not thinking about AI like it’s technology. They're thinking about AI like it's something different and special when actually it is just another technical workload.

It should sit in the environment that you've already secured for your web applications, mobile applications, CRM, and databases that hold all your company data. You've already got people securing it, controlling it, and governing it. This technology should not be different.

But you can implement detection mechanisms to tell if somebody's using it inaccurately or if it's grabbing information it shouldn't. This is not a web application on the internet. This is an enterprise architecture that you're building that includes all the elements of a secure safe system in your company.

Hoffman: What is one foundational concept every employee needs to master today regardless of where they sit in the organization?

Russell: The number one thing I encourage employees to do is to learn by doing. Don’t go to a model and ask it to solve your problem without understanding the problem you're asking. I call that vending machine AI.

That's how people use it: rewrite this, publish this, change this deck. Don't use AI like a vending machine.

Instead, come up with the problem and ask AI the way a producer would produce a movie.

I’m going to ask it what I want, how I want it, what I want us to think about, and how I want us to think about it. I'm going ask it to play certain roles. I want it to be contrarian. I want it to challenge my ideas.

Everything is bound by the questions you choose to ask.

Maybe you say, "I'm trying to get my customer to solve this problem, how can I do that?”

But you’re not saying, “Give me this PowerPoint deck.”

What if a PowerPoint deck isn't actually the best thing you could do for them? What if it teaches you how to build a little demo. Let your AI system meet you where you are but also complement what you're trying to do.

Hoffman: AI has become very easy to access. How do you believe the conversational nature of AI has changed the barrier to entry for non-technical workers and leaders?

Russell: People are in vending machine mode where they are defaulting to the truth of an AI model. But that model has no sense of truth, humanity, goodness, or right and wrong.

The model needs to know who you are as a company, your core values, your leadership principles, and your non-negotiables.

If those non-negotiables are leading the conversation, and if the AI is trained on them, then your AI initiatives will lead to business alignment. I always ask my business owners why did you start this company? Over the course of hitting ten million in revenue, or a hundred million in revenue, they cobbled together technology that limited the ultimate dream they had. Now all of that's worth rethinking.

So go back to basics. What are you trying to do? Who are you trying to do it for? Does it move the needle in way that’s aligned with your business? And most importantly, how are you going to measure it?

Then ask AI what is the best opportunity for you to deliver this with the expertise you have.

As Russell sees it, successful AI adoption starts with asking better questions. Organizations that build AI literacy, establish governance early, and align AI initiatives to business goals will be best positioned to turn experimentation into measurable results.

To receive future interviews and executive insights on AI, security, and technology leadership, subscribe to the monthly Perspectives by Splunk newsletter.