Navigating the AI Hype Cycle by Setting Realistic Expectations for Success

Artificial Intelligence (AI) is poised to reshape organizations in every manner. Setting realistic expectations for what’s possible with AI and how effective implementations work are important to help leaders and team members clearly understand its capabilities and prioritize initiatives that align with business objectives. All of this contributes to effectively navigating the AI hype cycle and ensuring your company properly reaches long-term goals. 

Recently, Andrew Simmons, Wavicle’s Retail Practice Lead, sat down with Ed Rybicki, Chief Information and Technology Officer at Mastronardi Produce, for an in-depth discussion about how organizations can practically leverage AI to get tangible benefits and drive long-term success at the CDO Magazine Summit in Milwaukee, Wisconsin. This conversation covered identifying AI use cases, setting the proper infrastructure, finding the right talent, and avoiding the common pitfalls that disrupt AI initiatives.   

Keep reading to explore the speakers’ perspectives on navigating the AI hype cycle and more or listen to the recording here: 

Andrew: Thanks everybody for coming. My name is Andrew Simmons, and I look after retail and CPG at Wavicle Data Solutions, a data, analytics, and AI company. I have been in data and analytics my whole career. 

I’m happy to facilitate an intimate conversation here with Ed, who is the Chief Information and Technology Officer at Mastronardi Produce. If you’re not familiar with Mastronardi Produce, you may be familiar with some of their brands. You can find Sunset Grown at Costco, Whole Foods, and Trader Joe’s, where I find their delicious sweet peppers. It’s a 70-year-old company that pioneered greenhouse growing. I’m sure Ed could talk about that if anyone’s interested; I know I sure was. It’s a $3.5 billion revenue company and a house of fresh produce growing and distribution brands throughout North America.  

Ed won the prestigious Chicago CIO of the Year award in 2021, among other numerous awards. He’s got 25-30 years of experience across IT, multi-disciplinary, and he’s been in the c-suite leading teams for quite some time, formerly at Vyaire and now at Mastronardi.  

There have been a lot of conversations today about AI, which I choose to think is not exhausting but a little invigorating. I mean, there’s a lot of energy around this topic. As Ed and I were trying to prepare for this, particularly on the back of what we’d heard over the course of today, one thing that we both thought was kind of unique about our session is that a lot of the chats today have been “outside-in” with industrywide perspective coming in, best practices, and what you should do in your organization. I’m excited to have Ed talk a little bit more “inside-out.”  

Ed’s living this battle day to day as AI is rolling out, and he’s at the intersection of a lot of converging forces. I’m sure he’ll talk about legal, business pushes, IT wanting to do cool and new things, consultants talking to you all the time about where they should go, and that type of stuff. I think we have a cool opportunity to learn from the frontline from the executive view of what’s going on in AI. 

Ed, I’ll start with identifying the right AI use cases and building a bridge to your business partners. Obviously, this is a critical topic for anyone in your role for all of your portfolio of technology responsibilities. But with AI, I think we heard over the course of today that the good news is the demand funnel is really stuffed. Your business partners in your organization are excited about AI; what that topic really means to them is something interesting.  

How does Mastronardi navigate balancing business interests with IT perspectives, ensuring the right pace and laying proper foundations for AI?  

Ed: Look, starting out, I think you know some of the other presentations hit the nail on the head. We’re all navigating new, unchartered waters. But, from where I sit, it doesn’t appear to be so different from technological solutions of the past. Not to mean AI isn’t different, but how you navigate that water isn’t necessarily different.  

If any of you have seen what is now a Gartner model – the hype cycle model – it usually shows that when something’s introduced there’s a huge amount of hype around it and the expectations are highly inflated. Then, you go through the trough of disillusionment, where now you get to the reality, and I think that’s a little bit where the AI cycle is. Although I think the plateau will be much higher than it has been for some other technologies in the past. So, when we’re looking at it today inside our company, there is some of that expectation.  

I was telling Andrew a story. We were at a fairly recent executive meeting where the top 40 people in the company come together every quarter for a couple days to talk through strategy, and I had a small AI presentation just to kind of get out and try to capture what I’ll call the leader position, because if any of you are practitioners inside it, sometimes it’s a weird thing where other c-suite or other executives are talking about a tech thing and they don’t immediately always look toward the IT department to give their weight on that, even though we’re the internal experts.  

To me, it’s about just understanding where the opportunities are, where the value is, not unlike any other piece of technology, but also being able to pivot and approach it in a way that may be not the traditional waterfall model or high analysis.  

I think there are a lot of things in the AI industry, not only single AI-based solutions but a lot of what I’ll call embedded solutions, like Microsoft Copilot. For example, I was just typing an email and Copilot kept coming over to the top of me saying, “do you need help with this?” I couldn’t figure out if it was just trying to help me or telling me that my email was really poorly written. I don’t know what it was saying. But you know those embedded things are already coming at us. 

Finding the opportunities to quickly start to use and test those and even get them in people’s hands is not disruptive in my mind. It can already be a helper and a little bit of a gateway into broader AI solutions, like maybe analytics using AI or something like that. End of the day, I would say, try to get yourselves in a leader position because nobody else knows this and the consultants are going to be knocking down the doors of your CMOs and your CFOs pretty quickly.  

If you can establish yourselves in that position that’s helpful, try to make sure that you understand where it can be helpful and where you maybe have to look at a standalone AI versus something that’s already embedded, but you have the opportunity to highlight both. Even with the stuff that’s already being embedded in the products you use every day, take the opportunity to highlight how that’s improving things, and it’ll contribute to credentialize you as an internal IT person.  

What kind of conversations are taking place with your executive leadership on AI adoption? How do they approach you to address their concerns and gain insights into AI development? 

Ed: We actually did have an instance where our Chief Legal Officer reached out, and he must have been talking with some outside legal counsel on the topic because the reach out was saying we have to ban AI. And I said, “okay, hold on a second. What did you hear and why are you hearing it?” What he was getting to was more about the open use of a Chat GPT or something like that, and the ability of data to leak back out.  

I think a lot of us, and a lot of the people in our companies, don’t quite understand that if they were to put a financial document or a document with customer information and say, “help me organize this better,” that some of those language models are ingesting and learning and potentially using your data. It was more about the protection of data and how to pivot.  

We came back with some guidance on the tools that we’re going to allow versus ones we won’t allow, and how to properly use these things. My education to him was that we can’t ban AI because it’s already here and can’t control it. I’ll go back to Copilot. It’s my favorite example because it’s almost impossible to turn it off.  

Andrew: It just turns itself back on. 

Ed: Yeah, it understands. Copilot’s privacy policy says it’ll not use your data, and that gave some comfort to our Chief Legal Officer. It’s part education and another opportunity to get in front because if my team wasn’t ready with a drafted policy, the lawyers were coming. Our CHRO was at a conference last week, and she took a picture of some slide like this and asked if we have a policy. She only started with the company a month ago, and I said, “yes.” So, it’s an opportunity to bring them together, work more in partnership, and tackle it together. 

What did you tell the leadership team during the top 40 leader meeting about your perspective on AI? 

Ed: A couple things. My first slide explained about AI and dispelled the myths. Next, I showed kind of a futuristic robot helping. Then I showed the Terminator, and I said that it’s the span of these things.   

But the way we kind of couch is that it is coming and there are huge opportunities for us to use it, but it’s going to be hitting us in a lot of different ways with everything from things that are embedded to generative AI for our marketing team.  

We’re the ones in the company that have the most knowledge about it, and it gives us a better opportunity to partner in those areas of the business. If it’s marketing, we talk to them about the different uses and applications there, and if it’s obviously our own department, we’re looking at how to exploit it. If it’s our greenhouse, we have robots and other technology that already has AI embedded in it to see if it can enhance our business operations.  

I think the overall message was that it’s coming. But we’re going to be your guide to get you through this; don’t be too afraid, but also don’t be on the top end of the hype cycle where you think it’s going to change the world overnight because, as we’ve played around with certain examples, there’s still a lot of work to be done to get it to be a true gamechanger. 

How is your corporate IT team exploring code generation and QA, especially with the productivity aids and features in AI and related technologies? 

Ed: A lot of this goes back to the platforms you’re in, and as we’ve moved throughout the time of IT, we’ve gone from these single solutions or custom solutions to now where we can buy a solution. But you’re still installing it somewhere that you own the servers or hardware or in the cloud.  

We’re anchoring on certain platforms, for example in internal IT, we’ve implemented Zendesk as our service management platform. That has different AI capabilities we can pick and plug in. Some of it is for checking code and code generation. We’re even starting to create a bot for our users that creates IT training documentation. 

We’re testing a chatbot for our warehouse, where people loading different products on trucks can quickly converse where to unload the products and it can respond with suggestions.  

There are environments where the workforce is transient — people that start in the morning and quit in the evening. The ease of getting them to adopt certain things is key, and we’ve always struggled with that in IT. These systems are complex, they’re not like picking up an iPhone yet. So, the AI we see can help us there as well.  

Andrew: That makes sense, particularly on the data and analytics side that we talked about forcing into the future.  

What key infrastructure elements are you working on right now to achieve AI success within the next 18-to-24 months? 

Ed: A cloud-first approach, and you need a platform of clean data. Most AI solutions are going to have something to do with information. Generative AI is different, and I’m not talking about how that would embed in products because I’m not in that space. But for corporate business use, a lot of it’s going to be based on data. If it doesn’t have clean data in a good spot, then you’re sunk.  

When you look at an AI solution that relies on the computational power of aggregated data and you don’t have those two things in place, you’re going to build a single silo for that solution or you’re going to tell somebody that’s excited that you can’t build the fifth floor without a basement. That’s where our push is now in my company.  

We’re going through a big modernization of technology. Not surprisingly, the agriculture world is not the epitome of technology. But greenhouse agriculture is because it is more of a technological-based way of growing. Now the corporate side has been a little left behind, so we’re modernizing a lot of that.  

So, you must be in the cloud and within certain ecosystems, whether that’s AWS, Azure, or even Salesforce. My suggestion is to not segment 100% of your stuff in an ecosystem. Companies are coming out and advancing their tool sets so rapidly that if you’re half in half out, you’re not going to take advantage of what’s coming. The other point is if you are in and have your workloads and data already in there, you’re primed and ready for when that next generation of tools comes out, whether it is Microsoft, AWS, or Google. You are ready to get speed to value if you’re already there. 

Andrew: Yes, it makes sense.  

How are you finding and retaining talent in the hot and emerging market for building and maintaining AI infrastructure? 

Ed: There isn’t a massive talent pool of AI-ready people because we all are learning it, and the ones that have already crafted that in their career are probably already sucked up by companies like AWS and Google. This leaves less for companies like us, but also it leaves an opportunity for us to build our people.  

When I’m running an IT organization, and my own people are a little bit apprehensive about changing their skill, I’ll say that they are in wrong industry discipline because in IT, you’re going to change constantly. You find that people pivot a lot easier when you explain its beneficial for them as well as the company 

When you have a core capability that can be insourced and built, you can augment it with consultants to help you along that journey. But I never want to be dependent wholly on that. We are building, training, and encouraging our people to get out there and start to adapt. Our support is in the form of time or dollars or other things to make sure they understand how these new technologies can work both for the company and their careers.  

Andrew: Makes sense. Well, we covered a lot of ground.  

Looking back at the past 18 months of rapid progress, what advice would you give others regarding navigating the AI landscape? 

Ed: The expectation gap is one of the greatest ones we’re going to have to watch out for — managing the expectations of what it can do and can’t do. Second is to get yourself in position to be successful with AI, which is unlike other software or hardware capabilities that have gone through the hype cycle.  

A couple years ago, robotic process automation (RPA) came on like gangbusters to the point where I literally bought stocks and then I watched it just tank. But I don’t see that with AI. I see it taking a similar path as cloud, because you could have said that about cloud 10 years ago, and now if you’re starting an IT department or company, you wouldn’t think of not running your stuff in the cloud.

I don’t want to get into a crazy philosophical conversation, but I do think it’s going to be a game-changing technology at some point as it continues to grow. From a business standpoint, human beings don’t love to change that rapidly. If you can show value and understand those use cases, there’s always a ton of opportunity. Once you’ve unlocked this, AI is going to continue to expand and develop.  

If you are in a leader position, don’t fall for the overhype. Don’t think it’s easy, don’t think that you can pay somebody one time to come in and they’ll get you all ready. Get yourself foundationally ready. If you’re going to bet, it’s going to be here for the long term and it’s just going to continue expanding. Companies that are in early and ready for it are going to have a big competitive advantage against ones that are sitting on the sidelines or not ready for it. 

Andrew: Thanks, I appreciate you sharing your views. Join me in giving a round of applause for Ed. Thank you so much for your questions and attention, I appreciate it.  

Ed: Thanks for joining.