How American University's Kogod School Of Business Is Redefining AI Education And Business Strategy

[00:00:04] - [Speaker 0]
How do you move beyond AI experimentation and actually prove it's delivering value inside a business? Well, my guest today has somewhat of a unique perspective on this question. Her name's Angela Virtu, and she is a professor of IT and analytics at American University's Kogod School of Business, where she also serves as the AI instruction faculty fellow. And not only that, she's also helped lead one of the first large scale AI transformations in higher education, while also bringing hands on experience from building AI solutions inside fast growing SaaS companies. So she's been on both sides of the fence here.

[00:00:51] - [Speaker 0]
But in our conversation today, I wanna learn more about why Angela believes that 2026 feels like a turning point for AI adoption, why so many organizations are stuck in what she calls pilot purgatory, and why focusing on real business problems and measurable outcomes is the only way forward. And we'll also explore the shift towards vertical AI, the growing importance of governance, and how leaders listening can balance innovation with trust in a rapidly changing landscape. Partners like NordLayer make it possible for me to attend events, speak with industry leaders, and bring those insights right back to you here on every podcast, every episode on the Tech Talks Network. I was recently at a tech conference where Mark Templeton said that the browser is now the computer. It was a modern take on the old idea from Sun Microsystems where they said many years ago that the network is the computer.

[00:01:51] - [Speaker 0]
And I think this perfectly captures where we are today. The browser is now the computer, and work has shifted into the browser, which means security has to follow. And this is exactly what NordLayer is doing with its new business browser. Instead of protecting the edges and hoping for the best, it secures the place where people actually do most of their work. And it also gives their company a better visibility, stronger control, and a more practical way to manage risk.

[00:02:23] - [Speaker 0]
It's one of those ideas that feels so obvious once you hear it. But if you wanna know exactly what that shift looks like in practice, please pop over to nordlayer.com/browser to find all the information you need. And please come back to me. Let me know your thoughts on it. But now, on with today's show.

[00:02:40] - [Speaker 0]
Let me introduce you to Angela right now. So thank you for joining me on the podcast today. Can you tell everyone listening a little about who you are and what you do?

[00:02:52] - [Speaker 1]
Hi, Neil. Thanks for having me. So my name's Angela Virtu. I'm a professor at the Kogod School of Business where I teach artificial intelligence. And I also am in charge of our culture building with AI.

[00:03:05] - [Speaker 0]
Well, it's a pleasure to have you join me today because I really love taking a a unique angle at the whole AI conversation. There's no avoiding it right now. And you're someone that's worked both in academia and building AI solutions inside fast growing SaaS companies. I've got to ask, how is that combination of hands on industry experience and teaching? How have all these things collided almost to shape the way that you think about AI adoption in businesses around the world today?

[00:03:32] - [Speaker 0]
Tell me about that.

[00:03:34] - [Speaker 1]
Before I joined academia, as you mentioned, I actually worked in a bunch of tech startups where I was building all of those AI solutions, both internally for our own operations and also externally for our our customers to be using. And so I think that experience really gave me the deep technical foundations of, like, how AI works, whether its existing limitations, and then that kinda translated really well into the academia world. And I think the biggest thing that's helped me bridge that experience together is coming up with a very solution oriented mindset, Where in industry, all of my AI solutions had to be solving some kind of business problem, and we had to be able to measure that with an ROI or a tangible economic reason. Right? We we love to build things just to build things, but unless it's actually solving a problem, there's really no point in doing that.

[00:04:20] - [Speaker 1]
And so I think having that problem solution orientation, really directed me a really nice transition into academia.

[00:04:29] - [Speaker 0]
It's so refreshing to hear you talk like that, talking about problem first. And I often feel sorry for students now because they're almost in education, training and waiting and learning for jobs that might not even exist yet. But as the, AI instruction faculty fellow at American University's Kogod School of Business, One of the things that really stood out to me and why I was excited to get you on today is you helped guide one of the first large scale AI transformations in higher education. So tell me more about that. What what did the process involve, and any lessons from that experience that that could even apply to organizations trying to introduce AI across their own workforce?

[00:05:10] - [Speaker 0]
What did you learn here?

[00:05:12] - [Speaker 1]
Oh, our process really started three years ago where our dean would bring these c suite or executives onto our campus to go and talk to students about what their job looks like, what they think is kinda going on in the workforce. And so we brought Brad Smith from Microsoft. We brought Brett Wilson, who is from Swift Ventures, which is a VC startup organization. And both of those individuals came to our campus saying, hey. AI is coming.

[00:05:38] - [Speaker 1]
It's here. It's as important as electricity or fire. And even in while we're listening, we might be a little like, maybe that's a little over overdramatic here. But even if it's 1% correct, like, it's gonna have really profound impacts on on the workforce. And as you mentioned in in education and then specifically in higher education, you know, our goal is is workforce preparedness for our students.

[00:06:01] - [Speaker 1]
We need to give them jobs. And so we immediately were like, well, we need to figure out what the expectation shift is because we can't be preparing our students for a world that doesn't exist. So we brought experts across all of our disciplines, finance, marketing, tech, brought them onto our campus, and we gave lots of training to our faculty about what AI is, how the existing industries are using this technology, what their expectation for our students are. And we really kind of started to ground the conversation, and here's what our students need, and here's what we're gonna need to do to change in order to help our students meet those those outcomes. Fast forward two, three years, Bloomberg's called us the first AI first business school, and that really comes down to the culture change that I've really been able to to push throughout our organization.

[00:06:51] - [Speaker 1]
Faculty, I love them, but we are some of the most hardheaded, resistant to change people you will ever meet. We're very opinionated. We think the way that we've been doing things is the way it should still be done. And so when you bring in artificial intelligence, what it really kind of did was it started to to bubble up all of these existing problems that had existed in academia for the past decades. Right?

[00:07:15] - [Speaker 1]
Things like academic integrity, other other issues about, like, lectures aren't really necessarily the best way to be teaching, things of that nature that we've kind of been able to sweep under the rug just because we haven't had that big disruptor moment. AI comes in, and now we can't really hide any of those problems anymore. So how do we get people who are super, super resistant and bullish and not wanting to change and kind of have them start changing the way that they think about it? The biggest thing that's been super helpful in our culture change is having a leader who's super bullish on AI, super supportive, and understands that everything that we try might not work. So we've been able to create this culture of experimentation through gatherings and just willingness to experiment that has really been able to to propel us forward.

[00:08:01] - [Speaker 0]
It's interesting how you mentioned that as adults and the institution and and education at large, there's this reputation for being almost resistant to change or very cautious around change maybe, and for good reason in some circles too. But what about the students? Are they less cautious? Are they really excited or feel feel overwhelmed by everything? How do the students feel about everything?

[00:08:24] - [Speaker 0]
And are are they excited now? Are they always been? Well, what have you seen from the students there?

[00:08:30] - [Speaker 1]
Oh, I would say that the students kinda have a very similar spectrum as their faculty. Right? You have 10% who are at the the cutting edge. They're gonna use AI even if we tried to ban it. Right?

[00:08:40] - [Speaker 1]
They're just gonna use it no matter what. You then have kind of that middle 80% who are at varying degrees of, like, okay. If we give them enough poking and prodding and trainings and exercises, they'll dabble and learn and experiment. And then you have that bottom 10% who are just not gonna wanna do it or, you know, due to environmental reasons or ethics or other, you know, beliefs that are, you know, not unfounded are just gonna be really, really resistant to that change. So I see, you know, yes, the students might have a little bit more anxiety about trying to find that placement since that bottom bottom rung of the career ladder is kind of starting to drop off.

[00:09:17] - [Speaker 1]
But a lot of it is is very comfortable to what we see with our faculty.

[00:09:21] - [Speaker 0]
It's really interesting how that mix of emotion spreads across every generation there. And in the workplace, I think many organizations have spent the last few years experimenting with generative AI tools, and then we've seen those problems with ROI. What what happens when you don't focus on the problem first and go tech first? Seen a lot of, projects stuck in pilot purgatory. So from what you're seeing now in 2026, do you think we're at a turning point?

[00:09:48] - [Speaker 0]
There was a lot of talk of AgenciKi and agents, but are companies moving from experimentation to real operational deployment to solving real problems? What what are you seeing here?

[00:10:00] - [Speaker 1]
I think we're at a really critical point in 2026 where if we kinda look at that tech hype cycle, we're kind of entering that trough of disillusionment. Right? We spent millions of dollars in all these AI pilots, and some people are starting to see those gains. Others, not so much. And so one area where I think we're really starting to see that that push from experimentation to operation is within our software engineering and coding world.

[00:10:23] - [Speaker 1]
Right? The whole role has shifted from being I need to go write and generate my own code to having AI basically supercharge me and become a 10 x or a 100 x engineer. So all of the big AI labs right now are basically making these bold claims saying, hey. A 100% of our code is basically AI generated, and we're now just helping orchestrate those AI systems. Where I now see that going is that we're gonna see a really big push, and we're just already seeing it right now, in verticalized AI.

[00:10:51] - [Speaker 1]
So AI that's specifically designed for lawyers, AI that's specifically designed for sales and marketing and all of the other specialized industries. And I think that's where we're now gonna be able to see that operational deployment where we're gonna have some best practices for each of those fields as to how to deploy, maintain, and operationalize those AI systems at scale.

[00:11:12] - [Speaker 0]
And looking at your career here, you're someone that's been closely watching the technical evolution of AI models. So if we look under the hood or sneak behind the curtain at some of the tech that makes all this possible, what what kind of developments in model architecture capabilities or tooling do do you think will have the biggest impact on how businesses actually use AI in the workplace over the next few years? Anything that you're seeing here that excites you?

[00:11:39] - [Speaker 1]
So there's two developments that I'm closely watching and following pretty closely. The first one is gonna have much more immediate impacts in the short term, and that's gonna be the move towards orchestration of these models. So a lot of individuals used to think that you can take a prompt, you can take some tools, you can take some data, and together, those three components are gonna give you your AI results that you need. Now we have a much bigger push into kind of the file tree system where we can define existing workflows almost like a folder where you have, like, one folder that's a workflow. Beneath that folder, that's where you can get each of those tasks, and each of those tasks will have your prompt or tools and your data.

[00:12:18] - [Speaker 1]
That orchestration is going to be much more resistant to all of the changes from the big AI company models that'll allow you to have much more robust implementation structures. The second thing that I'm really looking into is this push from just pure scale of these LLMs into the physical model space that's being developed right now. These physical models, instead of just being trained on lots of corpus of text, is actually looking towards how the physical world works. So think of videos of how I make a coffee or how I fold T shirts or, you know, more of those physical type jobs. And why I'm really bullish on those types of models is because I really think that that's gonna lead us to AGI, and that's gonna be the the baseline data that we need in order to have the AI transformation push into the more labor space.

[00:13:11] - [Speaker 1]
Right? So think of the robotics, how to make a coffee, baristas that are just gonna be making your coffee as a machine versus necessarily a whole person. And so I think the the interesting thing about that physical model is that when you think about how AI was talked about a decade ago, it was all about the replacement of those blue collar jobs instead of the knowledge thinkers that we're seeing right now with those LLMs. And the data that we're gonna be able to get through those physical models will really start to push the robotics space forward.

[00:13:42] - [Speaker 0]
And another big thing, but I think we're all seeing around the world right now is the economics of AI. And, it's interesting. Even though we're talking about the latest tech trend, the next big thing, the new shiny tools, I think that belts and, braces approaches to IT of, hey. You can only measure you can only improve what you measure is more important than than ever. So when leaders talk about the return on investment from AI initiatives, What is it that they should be really measuring to understand whether their investments are working?

[00:14:14] - [Speaker 0]
Because it it's something that was forgotten about when AI first came onto the scene. But what are you seeing now? What are organizations measuring, or what should they be measuring?

[00:14:23] - [Speaker 1]
So I think my most frustrating metric and what we really need to move past is just this adoption metric. Right? Like, that is the thing that is going to be empowering the rest of your workforce forward, but I think it's a really bad metric of assessing how well this implementation's going. Right?

[00:14:37] - [Speaker 0]
Yeah.

[00:14:38] - [Speaker 1]
So beyond that, it's gonna be a little bit industry specific. So right now, all of big tech, they're only talking about revenue per employee. I think there's a really viral clip right now of Jensen Huang basically claiming how he wants his top engineers to be spending $250,000 a month on just tokens, and that's how he's gonna be measuring all of his employees' effectiveness, which sounds outlandish when you think about how much the token cost is going to be outpacing people's salaries at a certain point. So I really like to ground this back to kind of how we started this conversation, which is just what is the problem that we're trying to be putting this AI solution into? And then what are the existing KPIs?

[00:15:17] - [Speaker 1]
So let's take a really, really tangible example here. And let's say we wanna have some kind of AI system to help with our customer success or customer questions. Right? Our support. We already have some KPIs that are gonna say what's our uptime, what's our resolution time on average, how often are we correct, right, or, like, what's our success rate in terms of those resolutions.

[00:15:39] - [Speaker 1]
You can then split into an AB test and say, here's half of our people who are gonna go through our traditional support system. The other half are gonna get an AI agent where we say, hey. I'm blah blah blah. The AI the AI agent. Right?

[00:15:50] - [Speaker 1]
How can I help you? You can use those same KPIs that we're measuring on this problem before on the AI systems to see how much better or how much worse that AI implementation's actually having on that problem. So I think the more grounded we can make those those AI implementations in those KPIs and in that impact, the stronger you're gonna be able to speak to that.

[00:16:13] - [Speaker 0]
100% with you. And I think if we look past the excitement and success of any AI project, IT will be waiting in the wings, wanting to talk about governance, becoming a big part of the conversation more and more now. So why are you seeing so many companies or are you seeing more companies beginning to prioritize AI governance? And and what does responsible oversight actually look like inside a modern organization? I'm sure if I was to ask five people, I'd get five different answers, but what do you see here?

[00:16:44] - [Speaker 1]
So I think the most successful companies that I've seen are using a distributed AI governance system, and I kinda wanna make a parallel to how cybersecurity, systems are already set up. So if you think about cybersecurity, you have one centralized team who's in charge of the response plans when incidences occur. They're responsible for the training as to how to tell all of your employees to look for a phishing email, things of that nature. So it's the responsibility of the centralized group to set the standards, to set the incident response. However, it's the responsibility of every individual within that company to be responsible, to be, you know, cybersecurity aware, not to click on these rogue emails.

[00:17:22] - [Speaker 1]
Right? Companies that are doing that very same approach within their AI systems are seeing a little bit bigger gain because they get that culture buy in a little bit easier. Right? You have one centralized group who can kind of ensure that the ethos and the security and the regulation is in alignment with that company brand and culture. And then it's the distribution out of all the individuals within that company to be able to use that AI responsibly and and well.

[00:17:50] - [Speaker 0]
I'm sure we'll have many business leaders listening to our conversation today who feel that pressure to adopt AI and do so quickly while still managing risks. So from your perspective, how can their their organizations better balance innovation with the controls needed to protect data, intellectual property, customer trust, and so much more. It's it's a real balancing act. But any tips or advice that you'd recommend around this?

[00:18:18] - [Speaker 1]
So the first tip that I would have is trying to create that closed system so that all of your proprietary information doesn't go beyond the the states, let's say, of that AI. That could be through an enterprise agreement where you basically have the rates where you say anything that we upload into here doesn't leave our our premises or it remains my intellectual property, or you could go about building it yourself with your own infrastructure. Obviously, that requires a lot more tech upfront costs, but would be another way to go. The second thing that I would say as you try to balance out that innovation with the control. Right?

[00:18:53] - [Speaker 1]
We can't just have entrepreneurship just for the sake of entrepreneurship. But, really, I think it comes down to the communication and the trust that you can build not only with your employees, but with your customers. So a really bad example of an AI implementation would be what Hilton has done recently, where they've replaced their booking agents from being live individuals, like humans, right, to being this AI system. However, if you picked up the phone and called them, it would say, alright. We'll connect you to a live agent soon, and then it's just an AI voice over.

[00:19:22] - [Speaker 1]
Right? Or it's an AI agent who's gonna be pretending to be a human. And they never really say that it's an AI that you're talking to, so people are going in with the expectation that this is a human. And then, you know, one or two minutes later when they can't really answer the question super well depending on how niche it is, You know, they find out it's just AI, and then they get really, really frustrated. So I think as you think about that balance, you really need to think about your brand and how you wanna communicate these experiments or these trials into this AI world and make sure that you're you're going with the right principles.

[00:19:58] - [Speaker 0]
And considering the pace of technological change that we've seen over the last three to five years, it's almost impossible to predict the future. But if I did ask you to look into a virtual crystal ball, where where do you see all this heading in the next few years? Any signals or trends that maybe business leaders should be paying attention attention to now if they wanna stay ahead of the next wave of AI development? There's always something else on the horizon. But anything you're keeping an eye on here or any anything you could predict over the next few years that you could see happening?

[00:20:30] - [Speaker 1]
The biggest thing right now for me is how you build trust within your brand, and that's gonna become increasingly more important as you have to cater to two separate audiences. Right? So the first audience is gonna be your humans, your actual customer base. How do they feel about this personal experience? How do they think and view about your company as a whole?

[00:20:49] - [Speaker 1]
But the second is we're gonna see this emergence of entire AI agents being able to buy and sell things. Right? So we're gonna have this new AI agent economy where you're now also not not only can sell to humans in real life in person, but also be able to appeal to how the AI, you know, the AGEO, the SEO for AI agents basically aligns. And those two things are really hard to thread the needle on because a lot of times, us humans like to be conveyed in a different language different tone or a different way than how those AI agents are gonna necessarily read it. So a few companies right now are starting to experiment on their website where they'll have an entire, like, call out box that basically says, hey.

[00:21:32] - [Speaker 1]
If you're an AI agent, click this thing. And then it's just, like, large markdown text file that, like, if you read it as a human, would be like, what is this? Right? But it's a way for the AI agents to kind of be able to read and parse through that data in that context a little bit easier. So I think figuring out that balance of trust on your core your core demographic and your core customer base is gonna be increasingly more important while you don't lose your own brand.

[00:21:56] - [Speaker 0]
Absolutely love that. And I'm gonna pull away my virtual crystal ball now and have a bit of fun with you. I'm not gonna pull out a virtual soapbox. Right? Because very often, I think we all see things online that frustrate us, myths and misconceptions around our area of expertise, and I'd love to try and boost some of those myths today.

[00:22:15] - [Speaker 0]
And one of the things I love about what you're doing here is you've got this unique vantage point as someone that's worked both in academia and building AI solutions inside fast growing organizations there. So when you've you've done all that, you've got all this experience, when you then start scrolling on Reddit or LinkedIn, you may see a few things that frustrate you. Are there any myths or misconceptions we can lay to rest today? The the the floor is yours. Go for it.

[00:22:40] - [Speaker 1]
Misconceptions on AI. Interesting. I think my biggest misconception is that AI is not necessarily the enemy, but it's the existing structures that aren't built for AI that is really to blame. And so I think AI has a really bad PR problem where it's kind of just the scapegoat of everyone's problems and issues and worries of that's the big catalyst. That's the disrupting force.

[00:23:05] - [Speaker 1]
And so I think AI has gotten this, like, bad PR rap of it's horrible for the environment. It's horrible for the economy. It's gonna destroy the entire world. But it's really what it's doing is disrupting the equilibrium of the existing structures and how organizations have been built. And until we can kind of take the 30,000 foot view and kind of ignore what we know and rebuild something from scratch, we're gonna keep having that friction between AI and how it gets implemented.

[00:23:33] - [Speaker 0]
Wow. Absolutely brilliant. A a real powerful moment to end our conversation on today. And for anyone listening, if we've I can hear the light bulb moments going off around the world. If they wanna continue the conversation or just look at more about the kind of work that you're doing there and follow you guys online, where would you like me to point everyone listening?

[00:23:53] - [Speaker 1]
So you can follow me on LinkedIn at Angela Virtu, and then you can follow my work at the Kogod School of Business on all social channels at Kogod Biz.

[00:24:02] - [Speaker 0]
Awesome. Well, I will add links to everything that you've mentioned there. It'd be great to get people to go check you out. I'd love to hear your thoughts on anything that you've heard today as well. I'm sure it's a a big conversation, start up on the economics and ROI of AI implementation to corporate pivots and AI governance and so much more in between.

[00:24:23] - [Speaker 0]
But just a big thank you for starting this conversation today.

[00:24:26] - [Speaker 1]
Thanks for having me, Neil.

[00:24:29] - [Speaker 0]
One of the things that stood out to me in our conversation today was Angela's focus on discipline and clarity. Yep. AI might be evolving quickly, but the fundamentals of business still apply. If you cannot tie it back to a real problem and a measurable outcome, it's very easy to get lost in the noise. And I think her perspective on governance also feels particularly timely because as AI continues moving deeper into core business operations, the organizations that succeed, they'll be the ones that treat governance as a shared responsibility rather than another centralized checkbox exercise.

[00:25:10] - [Speaker 0]
But I'm curious. Where does your organization sit right now? Are you still experimenting, or have you started to see real operational impact from AI? And what metrics are you using to prove that value? As always, love to hear your thoughts on this one.

[00:25:29] - [Speaker 0]
Please share your experiences with me. Pop by techtalksnetwork.com. You can leave me an audio message, send me a DM, connect with me on socials, or just browse through 4,000 interviews, or check out which events you might be able to see me at and meet me in person. Let's break the fourth wall. Let's make that happen.

[00:25:49] - [Speaker 0]
Other than that, big thank you for listening as always. I'll be back again tomorrow with another guest, but thanks for listening as always, I'll and speak to you tomorrow. Bye for now.