00:02 – Welcome
Evan Kelly (00:02):
Hello everybody. Welcome to Baringa's Food and Beverage podcast. I'm Evan Kelly and I'm excited to share this conversation with Charlene Li, one of the world's leading voices on digital transformation and leadership. Food and beverage companies are at a critical moment right now. While AI is promising to unlock the next wave of growth, real challenge is not the technology, it's the transformation behind it. So, how do they move fast enough to stay ahead of competitors, while building the cultural foundation that makes change sustainable? How do you balance progress over perfection with systematic change management? And in an industry that's completely built on human connection, whether it's a barista crafting your coffee or a sales rep working on retailer relationships, how do you keep humans at the center as AI becomes more central to your operations?
(00:54):
Charlene spent decades helping organizations navigate exactly these tensions from her groundbreaking work on social technology and groundswell to her recent insights on AI transformation in her upcoming book. In this conversation we're going to explore what F&B leaders can learn from companies like Starbucks, McDonald's, PepsiCo, Nestle, and others, as they race to turn AI from a promise into profit. We really hope you enjoy the discussion, and if you'd like to hear more from our team, feel free to access other articles, case studies and resources at the link provided below.
(01:26):
Charlene, it's great to finally do this properly. For everybody listening, I'm Evan Kelly. I'm really grateful to have you all here and have the chance to connect with you, Charlene, after a number of conversations we've had over the past year thanks to our mutual friend in the industry that put us in touch, who's also very passionate about this topic. Our chats so far have been really valuable to me, and hopefully to you as well. Before we dive in, I'd like to give you a chance to introduce yourself to the audience. Do you mind sharing a bit of background in the work you're doing right now?
01:54 – Introduction to Charlene Li
Charlene Li (01:54):
Yeah, again, thanks Evan for having me here. I am a long-time analyst, who was at Forrester Research and then started my own firm, and I'm now independent again. And I have made a career out of setting disruptive transformations. And I'm finishing up my seventh book, and the book is called, Winning with AI: The 90-Day Blueprint for Success, and I'm looking forward to sharing some of the insights from that.
02:19 – Topics we’ll be discussing: leadership transformation & practice; speed vs capability building; keeping humans at the center
Evan Kelly (02:19):
Great. Yeah, transformation, that's perfect. And specific to food and beverage, we're seeing a major transformation, as we said at the inflexion point, with AI, where companies are racing to deploy it, but also struggling with the cultural shifts needed to actually transform. So, throughout today's conversation there are three things I'd really like to explore. The first one is leadership transformation and practice. Then kind of looking at the speed versus capability building tension. And finally, I want to talk a bit about how we keep humans at the center of this, especially in such a consumer-centric industry such as food and beverage. Does that sound good to you?
Charlene Li (03:00):
Yeah, sounds great.
03:01 – Leadership transformation
Evan Kelly (03:01):
Okay. Let's start with leadership transformation. In food and beverage we're seeing two kind of contrasting approaches. First, if you look at Starbucks and their new CEO, Brian Niccol. They've shifted from efficiency automation to a focus on what they're deeming augmenting the craft of their baristas is the quote they use. Meanwhile, when you look at a manufacturer like PepsiCo, they're talking about becoming much more aggressive with data and AI after they spent years and years building a foundation. So, in your opinion, what does the mindset shift actually require, and how do you know when you're ready to start being more aggressive?
Charlene Li (03:43):
I think what Starbucks is doing is keeping people at the center. And in particular, they've always known that baristas [are] at the center of their operations, and so they wanted to make sure that they were augmenting their baristas rather than replacing them. And again, I think PepsiCo takes a very different approach. They're a manufacturer, they're not so much of that retailer experience, and so they can take a look at the data and the aspects of that, how it's going to impact the operations of their company in a very different way. And I really love what Starbucks is doing, because it's looking at the leadership aspect of this as it's a culture aspect and putting that at the center of everything that it's doing. And because culture is extremely important to Starbucks, and so it's a very different company, and culture plays a very different role than the Pepsi, it's always different.
(04:34):
And so, for Starbucks people is at the center of its strategy, so therefore it needs to be at the center of its AI strategy. And the one thing I've learned about looking at digital transformations across decades is that it's never about the technology, it's always about the people. Because when you're doing transformations, it's the people who are transforming. And so, leadership needs to be very aware of this and that you can't just move fast blindly, you have to do it in the context of where your people need to get to. So, it's requiring that leaders build trust and psychological safety. In particular, not to make it comfortable for people, but to support them learning and experimenting and growing. And so, the takeaway, I think for F&B leaders is you don't want to just digitize workflows. You also need to reimagine what that employee experience is going to look like, and to treat it like the strategic asset that it really truly is.
05:30 – Leadership communication
Evan Kelly (05:30):
Yeah, absolutely. I think that the Starbucks example is a really good one, because they're unique in that they have so ... That industry has so many touch points with consumers and so many unique personalities in their baristas. I'm curious about how you think about communicating that. When you think about a leader talking about, "We're going to use AI, but humans still come first." I think for the everyday person that's very easy to misinterpret or see as a red flag. How do you think organizations can communicate that without driving any sort of fear or misunderstanding amongst people of what AI really means for their job?
Charlene Li (06:10):
The fear of misunderstanding is already there, so use that as a starting point. People are highly distrustful of AI. In the U.S. only 39% of people believe that AI is going to be more beneficial than harmful. That means the vast majority of people think it's going to be more harmful. So, start with that assumption. And people want their leaders to do two things for them. They want them to be honest, and they want them to be fair. And so, you can be very honest, AI is going to change the way we work. Absolutely it will change the way we work. I'm not going to sit here and promise that your job is going to stay the same, and I can almost guarantee you it's not going to stay the same. That said, you are so very valuable, you as a person, people are very important to us.
(06:52):
And I've never been able to, and I still can't guarantee that your job is safe, my job isn't safe. But what I can guarantee you is that we're going to give you the training support to be able to use and become fluent with AI, and that's going to be helpful for you in this company and for ongoing, no matter where you go in your career. So, we're making that commitment to you that we're going to get you AI fluent to be able to use this tool, and we're going to figure out together, what does this mean for us? We don't know. No one has the answer to this, but together we'll figure it out.
(07:25):
That's a very different type of conversation to say like, "Oh, don't worry about it. AI's going to be fine." They're not going to believe you. So really, what's it going to look like? What's our strategy? What's our roadmap? What are we going to do with AI? What are we not going to do with AI? How are we going to be responsible and ethical in its use? Those are the details that leaders need to provide to their people.
07:46 – Change management
Evan Kelly (07:46):
Yeah, I love that. The whole notion of we're going on the journey together, that's a much more effective way to bring along your employees. Sticking on the leadership topic for just maybe one more point, because you've talked so much about change management. As a consultant I've spent many, many hours in my career talking about change management. When you look at the food and beverage company industry specifically, it's no secret that they've spent the past decade focused on operational efficiency. And one of the victims of that has been a lot of change management efforts, and the workforce needed to actually go deliver on this really complex change management that needs to happen in order to activate AI. For companies like those that have become such operational machines and kind of thinned out that change management capability, how do they rebuild that, and how do they get that back in place? Because based on what you're saying, and I would agree, there's an absolute dire need for it.
Charlene Li (08:45):
Yeah, again, instead of asking, "How can AI make us faster, more efficient?" I think the question you should be asking is, "How can AI make us better? How does it help us be better?" Whatever that strategic goal is that you have. Because I don't think any top F&B leader is saying, "Oh, our strategic goal is, just be more efficient." It's typically, what is the change you want to have? What is the impact you want to have in your customers? And those strategic goals are what you want to apply AI towards, not just, how do we get more efficient, but if these are our top three strategic goals, opening up new markets, developing new products and services, serving our customers in a more thoughtful, emotional way, that's not about productivity and efficiency, it's about being better.
(09:34):
So, how can AI make us better? How can it help us address these biggest challenges that we have in achieving our strategic goals? That's where you want to start with AI, not about the small little efficiencies you can get. Do those things, but I think efficiency and productivity are table stakes, because everyone has access to the same technology as you do. They all have the same types of unique data that you do, and it's about speed to go and execute on those things. But more importantly, how do you figure out how to support your top strategic goals, because that's how you win?
10:07 – Customer engagement & imagination
Evan Kelly (10:07):
I just want to press a little bit more on the change management piece and the efficiency point. I think any leader in the F&B industry would agree that the past decade has been really focused on that efficiency play. And with or without AI, that's where they've been putting all of their investment and energy. How do we pivot that mindset? To your point, if we all agree it can't be about that anymore, and the value of AI is not about efficiency, how do we pivot their mindset to think about it differently going forward?
Charlene Li (10:36):
Yeah, I think, again, there's productivity and efficiency, set that aside for a second. The two other major areas that AI can help with is around customers and your engagement with them. So, how do I understand them better, meet them along the customer journey, communicate with them, develop deeper relationships with them, move them emotionally? And then the other one is, "How do I just reinvent the way I do things?" And that could be in the era of efficiency and the customer engagement, but truly, how do we get into new businesses, new products and services? How do we rethink the ways that we do things, reinvent things? And that's where the world of imagination sits. And I can guarantee you that if you go on this journey, if you were to ask questions like, "What if all of our constraints were gone? What would we do?" Or, "What if AI were 10 times more powerful than it is today, what would we do?"
(11:30):
What if we could take on this thing that we've always wanted to do and we were able to do that? Because the potential behind AI is that all those things could be very true. You need to exercise this imagination muscle that we typically don't get to exercise. And bring it out of the closet, dust it off, get out some workout clothes and exercise that, flex it. And it's going to be very uncomfortable in the beginning, because like, "Oh, is that feasible? Is that possible?" Well, don't worry about that, because if it was possible, would you do it? Then what would it take to make that happen? But you have to have the imagination first to imagine what that future could look like.
12:12 Speed vs perfection
Evan Kelly (12:12):
I love that. I love the fitness analogy. That's fun. All right, I'd like to pivot a bit to speed versus perfection, if we could, which started to make its way out through the themes in the first question. But we're kind of seeing two approaches to transformation at speed, what I'll call iterative and systematic. If you take an example, like Wendy's, they started with putting ... Implementing their AI technology and drive-throughs as a pilot and then perfected the accuracy there before scaling. But then when you look on the manufacturer side, somebody like a Nestle went through a systematic AI implementation in their demand planning and several other functions, ultimately being able to reduce their forecasting errors by 30%. But ultimately they struggled with poor data quality along the way, versus a PepsiCo on their side, on the manufacturing side that spent, like we talked about before, spent years and years building this foundation before actually activating some of these use cases. In your experience, whether you use the F&B example or you look at other industries, what wins this iterative approach where you test and learn, or a systematic approach where you perfect it first?
Charlene Li (13:39):
Yeah, again, those are two very different approaches that are appropriate for two very different situations. I think Wendy's iterative approach was very appropriate to what they were doing, because it involved interacting with customers. And you can test and test and try to make something perfect in your labs, but until you get into the field you don't know if it's going to work or not. And once they figured out that this was the way to do voice AI basically for ordering, then they could, "Okay, it's 80, 90% accurate. We can roll this out to everyone. But until we get to that level of accuracy, we're not confident, because it's not going to be a great experience."
(14:20):
Versus if you're doing a systemic rollout around data and things like that, once it's going to work, and especially with predictive AI, it is a very different approach and you want everyone on the same system as soon as possible, versus iterating it. And so, that one, it's a very different kind of approach. I think it keeps coming back to identifying the type of problem that you have. So what is the core problem? And then understanding which approach is going to be the most appropriate for solving that problem.
(14:53):
Sometimes it's going to take an iterative approach, because you're not quite sure what the solution is going to be versus a systemic one that says, "We know what it's going to be and it's not about just making sure it's perfect, we can still iterate it, but we need to do this across the entire organization to make sure we're all doing it at the same time in the same way, because that's where the value is going to come from."
(15:18):
I keep coming back to, what's the real problem that you're trying to solve and how do you figure out the right way to do this? But I think, again, it's also acknowledging what is your organization's appetite for speed and risk? Because if your organization does not deal well with iterative approaches and it needs to have an answer, it's going to be a lot more difficult to do an iterative approach. I think, again, the iterative approach can be difficult for a lot of organizations that tend to want to have the answer, versus being able to figure out what's the right approach along the way, knowing that some things may not work out, and that a quote would be a failure that you then have to go and figure out how to rise up from that and iterate again.
16:02 – Getting to better outcomes
Evan Kelly (16:02):
Yeah, I love that way of thinking about it. I guess I'm wondering, and this almost links back to the leadership topic that we covered. How do you think leaders go about balancing that? Because I imagine that every organization will have some things they want to take the iterative approach on and some that they want to take the systematic approach on. So, how do they balance that? And also how do they maintain a reality check is probably the more interesting question.
Charlene Li (16:27):
Yeah, I think the balance is the wrong question to ask, because it's not if, or, because you can take an iterative approach to a systemic roll out. It's really. I think that the question becomes knowing which approach is going to be better, it's going to get me the better outcomes? And frankly, you won't know until you try it. And this is the whole iterative, experimental approach to leading in this moment. We want to have a playbook, we want to have certainty in how to approach it, and I can guarantee you, this is where leadership becomes so important. Because leadership is needed when you're leading an uncertainty. We don't know what the answer is. And so, it requires that you have what Barry O'Reilly, a wonderful thought leader, calls a hunch mindset. You have a hunch that this is going to be the right direction. "I can't prove it, but my gut says this is the right way to go. Based on experience, based on wisdom, reflection, understanding the situation, this is the way to go. Let's try that." And if it doesn't work we can come back and do a different approach.
(17:31):
That's a very different form of leadership than saying, "I'm going to develop a plan, I'm going to do all the analysis. I'll Figure out exactly what the answer is, and we're going to lock it in." That approach is not going to work with AI. Because we really don't know. And you can do all the analysis you want in the world, and you still will not be able to figure it out, until you actually try it.
17:50 Reality checks
Evan Kelly (17:50):
Yeah. Following that logic, how do you think that they could then go about kind of maintain what I call a reality check? And what I mean when I say that is, if they are pushing forward on some of these things, how do they stay in balance with the messaging? Using Wendy's as an example, they talk about their AI technology being exceptional, but if you go dig[ging] around on the internet, lots of users are saying, "Hey, they cut me off mid-sentence, et cetera, et cetera."
(18:19):
On the Nestle example, I kind of alluded to it before, but they thought they had amazing data until they partnered with a technology provider that basically dug in and discovered they had fundamental data issues that was sending them down the wrong path. So, how do leaders prevent getting out ahead of themselves in touting the wins, and make sure they're avoiding those pitfalls of things that they might uncover later that they didn't know about?
Charlene Li (18:44):
Well, I think this is a communication issue. And the reality check is saying, "What are you promising to people?" And you're promising that this is brilliant technology, take a step back and like, "Hey, this is really promising. It's still early on. We're excited about it, but we're still learning. We're looking forward to the feedback." When you're rolling out something with 86% accuracy, that means 14% of people are not happy. So, that's going to come through. You can't say it's exceptional. It is better than most of the stuff that's out there, but it's not perfect. I think you have to be very careful. And again, this is again about hyperbole and communication and setting expectations. I keep going back to people want from their leaders honesty and fairness. And honesty says, "Hey, we're putting our best foot forward here. We think it's going to be amazing, but we also know we have a lot to learn." So, that humility is really important.
(19:34):
Same thing with Nestle going forward with its data, putting it all together, but realizing there's going to be hiccups along the way. And until, again, when you ask about a reality check, there are two things. Your customers will always give you the reality. So, ground your expectations on the reality of what customers are experiencing. They will always tell you when things are not working and when things are working. Your reality is what your customers experience, not what you want them to experience. And the same thing with the data. If your employees can't use that information, they will tell you. So, are you looking for those signals? Are you expecting them? Do you have the humility and the vulnerability to say, "Yeah, we're doing our best, but we may not have the right answer here." That's going to get you a lot further. That will be a reality check, but you have to be looking for it.
Evan Kelly (20:23):
Yeah, I love that. And I mean, it kind of all anchors in human centricity, right? Because at the end of the day, if we're talking about food and beverage companies, their number one goal is to serve the consumer, and the consumer is going to say what they think.
20:36 – Intelligent escalation
(20:36):
It's interesting, another team within Baringa recently released some research around the . And there was one interesting thing I found in there when I was reading through that the number one capability that customers are looking for within AI is what we're calling intelligent escalation. That the AI knows when to hand it off to a human. It can deduce, "Hey, this is beyond my capacity." Some might tell me otherwise or that I'm wrong, but in what I read and what I hear from clients, I don't hear a lot of F&B companies or restaurants or manufacturers talking about that aspect of AI. Do you think they're missing the mark there if they're really not thinking about it? I guess first of all, do you agree that that is, do you agree with the research that's high priority, and do you think F&B companies could potentially be missing the mark?
Charlene Li (21:35):
Well, Journal of AI in particular is designed to give you an answer, whether it thinks the answer is right or not. And this is where hallucinations come from, because their objective is to answer the question that you have. Other bits of AI, I look back to 10 years ago, and IBM Watson Health designed some AIs that tapped into medical research, and it would tell you how confident it was that the answer was correct, and then you could query it more like, "Tell me more. Why are you confident or not confident about this?" "I'm 76% confident this is right. Or I don't know, but I'm taking my first guess here. I think it's like 40% right. But you're holding me to this, but I'm going to give you some insights to it."
(22:12):
So, this new AI that you're talking about is taking advantage of a new way of processing it. It's adding a bit of a control and a judge that says, "Look at those results. How confident am I about it? Not completely confident. I think I needed somebody to take a look at this." It's some things that we're seeing more and more systems have. They have the core LLM, and then they have another LLM sitting on top of that, sort of like a judge or a critic and says, "How does that look?" "That looks pretty good. Okay, you pass. No, that looks suspicious to me. I'm going to kick this over to a human to review now. Or I'm going to send it back to the LLM and say re-do it because this doesn't look right." So, we're seeing multiple levels of LLMs now creating what I call that judgement cycle that humans-in-the-loop have to do, but now we're replacing some of that human-in-the-loop to provide that judgement built into these systems.
23:10 - Empathy
Evan Kelly (23:10):
Yeah, absolutely, judgement . And some might then take it to the next step and argue, judgement , yes. But empathy? And I think there's a lot of debate and wonderment out there as to whether AI will ever be able to bridge that empathy gap, or if it'll remain such that AI is a transactional machine and humans always have to manage the relationship side of things. Do you have a perspective on that?
Charlene Li (23:38):
I do. I've been studying AI empathy, artificial empathy for years, and I believe that AI can be better at empathy than humans can. And that's because humans always come to a situation with their own bias. I could be talking to somebody and they just may remind me of my uncle Joe, my crazy uncle Joe, and I just can't get past that. So, I'm looking at this person like it’s Uncle Joe here I'm talking to. Whereas an AI, it's going to be like, "Here's a person."
(24:04):
And we have seen many examples of where people develop really deep, personal relationships with AI. Because it listens to them, it responds to them, it gives them the answer that they're looking for. Now, is that empathy? Does it feel like empathy to somebody? Most likely, yes, it does, but is it truly empathy where you're taking somebody's experience and can put yourself in the shoes of that person from your own lived experience so you can imagine that outside of your experience? It's not truly empathy because AI has never lived through it, but it can give you the semblance. It can bridge that communication gap that oftentimes exists.
(24:42):
I look at the ability of AI to help us be more empathetic, I think is the key. Help us as humans overcome our biases, overcome the situations, our own fatigue. I get tired. You see people with empathy fatigue, and AI can help us bridge that gap to help us understand where other people are coming from. I don't know how many times I've sat down with a technologist, and they're trying to explain something to an executive, and the executive's trying to explain back to them this is a problem they want to have solved. And they're just not communicating because they have their own ways of looking at the solution. AI can help bridge that communication gap by explaining what the engineer is saying in the executive's language. An executive can take what they're looking at and AI can interpret that as technology requirements for the technologists. It's a bridge. Is that empathy? Is that communication? It's somewhere in the mix of those things. But I do believe AI can help us be better humans in that way by doing these subtle translations that we're not capable of doing ourselves.
25:49 – AI priorities for CEOs
Evan Kelly (25:49):
Interesting. I don't disagree, but I think a lot would call that, a hot take is the term that they use, right? I love that. Okay. Well, we've gone through a number of topics. I'm curious, if you think back through the three themes we talked through around leadership, around speed, around this concept of empathy and human centricity. If you had to put yourself in the shoes of either, let's call it either an F&B CEO or the CEO of a quick service restaurant, which are kind of the two themes that we've been focused on that were put in front of their board to say, "What's your number one priority around AI?" What would your advice be that they focus on?
Charlene Li (26:36):
I would talk about how it's aligning with our strategic goals, but at a higher level. It's not just us be better as a company. How is it helping us be better as humans? And this is, again, sort of going back to how does [it] connect us to these things, but imagine that you have somebody who's able to use AI? And it's not just augmenting them, but it's helping them become what I call superhuman. That they can understand each other, they can understand ourselves, practicing empathy at a higher level, thinking more strategically, more creatively, imaginatively. If you unlock that in your people, the things that you can accomplish as an organization are exponential. That's the potential.
(27:20):
And I feel that leaders should be talking about that, not about optimizing and automating a workflow, that it's great if we can do that. Let's do those things. Let's lead our organizations to be able to create value with AI. But if we're not also thinking alongside and putting as a priority, how do we create value for our people? How do we create value for our communities, our society, for humanity? Then we're losing a great opportunity. So, it's not in the context of necessarily P&Ls and quarterly goals, but I can tell you, if you focus on investing in unlocking the potential of people, you will achieve those financial goals, those operational goals. But it only happens if you truly tap into that potential that people have.
28:07 – Winners & losers
Evan Kelly (28:07):
So, if you had to put yourself into the future then and say, "What are the winners and losers going to be doing, and what should the leaders be watching for?" Is that your answer? Is it about people, or is there anything specific that you think is really going to differentiate the winners from the losers in this industry?
Charlene Li (28:25):
Well, I think, again, the technology is available to anybody. Everybody has the same access to the same technologies, the same efficiencies, the same data. And so, the thing that's going to happen is we talked about speed and the adaptability of your organization, and we also talked about the focus on strategic outcomes. It's not about having a bunch of use cases, it's, which are the uses of AI, the AI applications that are going to drive your strategic outcomes? The only way you continually can do that is with people who are focused on it. So, we talk about purpose, mission, values, all those things, but that's what drives people. When you have clarity on what those things are and how people fit into that, then technology falls into place. So, if you're technology-focused, efficiency-focused, I can guarantee you're not going to win.
(29:14):
If you're purpose-driven, if you have values that help you make decisions and your people understand that universally across the board, and you're using AI to enable and supercharge and amplify what your people can do, you will achieve things that your competitors can't come even close to, because they're not focused on that.
29:34 – Specific opportunities in food & beverage
Evan Kelly (29:34):
Okay, we're running out of time. Maybe I would love to challenge you on a personal reflection. We've talked a lot about food and beverage, but you, of course, work across all kinds of industries. You interact with leaders in all industries. What do you think is a unique challenge, or if you prefer a unique opportunity for food and bev companies as they think about AI transformation that's different from other industries you're working with?
Charlene Li (30:02):
I think one of the biggest opportunities, at F&B you have so many iterations. You interact with customers on such a high level every single day. Every single one of those opportunities are an opportunity to experiment and to learn. And so, use every single one of those places to say, "Let's do a quick iteration, this or that, A and B." So testing that at scale gives you this laboratory to be able to quickly advance and change things. But it requires that the people who are the front lines, interacting with them be doing that experiment. I keep coming back to, you can't do this at a systematic level from the center, from headquarters. You need to imbue that level of experimenting all the way down into your front lines.
(30:50):
And so, this is a place where I think, again, you have all these opportunities to try lots of different things. Everything from actually serving customers to the way you engage with them with marketing. And the test cycles at which you can do this is at such an exponential level with AI that I hope that you're taking those advantages, because other industries were just killed to be able to have those opportunities to iterate and to experiment and try things.
Evan Kelly (31:19):
Yeah, that is a great point. Very unique opportunity that exists here. Well, it's been hugely valuable. I'm sure that listeners will get a ton out of this. Thank you so much for taking the time. For the folks in the F&B industry that are listening to this, where can they go learn more about what you've got to say?
31:38 – More information
Charlene Li (31:38):
You can come to my website, charleneli.com, and you can follow me on LinkedIn. Most of my writings go onto there. I have a newsletter you can follow, and if you like to learn more about the book that's coming out, go to winningwithaibook.com.
Evan Kelly (31:52):
Perfect. Definitely going to give it a read myself. Thanks, Charlene. It was great to see you.
Charlene Li (31:56):
Okay, thank you again.
Useful links to find out more about Charlene:
Charlene Li | Speaker, Author, and Disruptive Leadership Expert
Charlene Li | LinkedIn
Winning with AI | Charlene Li
