Data, ethics, and empathy: the future of technology at work

Claire Stewart: 0:00 Today’s LBIdea is to never make a decision that you can’t explain. Our guest today is Matt Brady. Matt is an award-winning assistant teaching professor in Leeds’ Organizational Leadership and Information Analytics Department, where he promotes building ethics and empathy into digital transformation and business technology. A serial entrepreneur, Matt’s latest startup is Volley Solutions, a collaborative decision-optimization platform that helps organizations make important decisions using information instead of intuition. Thanks for joining us today, Matt.

Matt Brady: 0:34 Hey, it’s great to be here, Claire. Thank you for having me.

Claire Stewart: 0:36 Well, we’d love to get into everything you do with students and everything you do in the startup world. But first, to start off, what originally sparked your interest in the intersection of technology and human behavior and ethics? We’re starting with a big one.

Matt Brady: 0:52 It’s a great question. Technology is only interesting for me if it helps people. At Leeds, we focus on what most people consider to be bleeding-edge technologies, though our researchers have been working in these technologies, in some cases, for decades. And I see from an entrepreneurship and industry perspective that the technologies are being used effectively. So there might be some risk or some unknowns. But it’s certainly my comfort zone to be on that bleeding or leading edge, working with students to prepare them as our future leaders. I have had a lifelong interest in technology. I studied first engineering at Purdue and then quickly found out that I was a better fit in the School of Technology. So that’s where I shifted my focus. And even dating back earlier, my mom was a teacher, my dad was in sales. So I’ve always been very human-centered in my approach to leveraging technology. More recently, I’ve gone through a series of startup businesses. Many of those have been sold or in some way acquired. And then I founded Volley Solutions, as you mentioned, in 2020. And more recently, a civic-oriented business called WorksBetter.

Claire Stewart: 2:07 Well, you brought up Volley Solutions. So I wanted to ask about that. You advocate for making data-informed decisions through that organization. So what do you see as the relationship between information and intuition in companies and organizations?

Matt Brady: 2:21 Well, I think they’re partners. I think you need a method for when they disagree. So I’ve become much more attuned to and respectful of gut and intuition as I’ve learned more about decision science. I think accomplished leaders, as I work with them speaking around the country, their instincts are proven. So the way I have taken insights from that exposure is that your gut is a dataset too. It’s just one that no one has audited. When the two disagree, that’s the cheapest research you’ll ever get. And so Volley, as a decision science platform, allows for executives to reconcile between how they feel a decision should be taken, the data that they’re receiving, the inputs from their leadership team and their entire staff, and the decisions are made more holistically and hopefully more effectively.

Claire Stewart: 3:20 That’s really interesting. It’s very cool stuff. So where do you see people commonly hitting roadblocks when it comes to decision-making based on this work?

Matt Brady: 3:28 Yeah. One of the major roadblocks I see is that every dataset answers somebody’s question. The key is to make sure it’s yours. And to that end, what we really work on, both in my speaking and in the Volley platform in business, are techniques that are most effective to address a culture of excellent decision-making. Those would include pre-mortems, red teaming, and what’s called “disagree and commit,” where there is time to share perspectives on said decisions. Ultimately, you have to seek out those naysayers. One of the things that I coach executives on is to look for those who are silent and who may have the most insight, where that information is just not being conveyed upward. It’s not being escalated.

Matt Brady: 4:19 And so the goal is to hear from those folks because they often, at the edges of their organization, have the best feel for what is happening and also have the ability to yield the greatest effect on the projects that are being funded and resourced. So technology here can either help or hurt. As we have all seen, especially with the advent of large language models and AI more broadly, technology can pull scattered data together. It can get you to the wrong answer faster, which can be concerning in this world of shorter attention spans and tighter deadlines.

Matt Brady: 4:59 Ultimately, my recommendation and my goal is to keep a human in the loop. I think that’s the most effective approach for all uses of technology.

Claire Stewart: 5:09 So you mentioned listening to naysayers in your organization and making sure there’s a human in the loop. Are there any other ways that you would say that teams could build a culture of decision optimization?

Matt Brady: 5:18 I do. I do. I think, obviously, broadly speaking, I coach my students to look for organizations that allow for risk-taking and celebrate failure, which sounds a little counterintuitive. But if you’re punished for failure, and in some industries that is necessary, let’s say, but in other industries, a risk-taking or experimentation model can be very effective in driving bigger and faster results. I think those are cues to allow for more innovative, more resilient organizations. I think the same can be said about decision-making. Naysayers should be recognized as valuable. They build greater pressure on the decisions that are made. They argue for stronger data and more compelling arguments. I think those are key.

Matt Brady: 6:08 As a technologist, though, I also have to say that because we have the ability now to reconcile gut with external datasets, with feedback from all of the stakeholders, including employees, customers, etc. We can get the “both and.” We can be more rigorous in the quality of our decision-making while bringing higher quantities of data in and hopefully have what would be considered best case scenario for leveraging both dimensions of inputs to yield better outputs.

Claire Stewart: 6:38 That’s great. So really, in your experience, technology is helping this process, not hurting it and separating us, really. It is really helping the decision-making process from the case studies you’ve been able to see?

Matt Brady: 6:48 Absolutely. I think that organizations, and we see this at the GDP level, at the yield, at the for-profit and, say, nonprofit level, even teams and individuals have higher output, more throughput, and hopefully better satisfaction along the way. I would say that it’s certainly a period of great change. And with that, we have to be cognizant of people’s ability to stay connected, to feel valued, to not get run over, but instead be able to leverage and harness these great capabilities. But my sense is it’s a very bright future. And I think that, if anything, AI has caused us to revisit some of these questions around whether it is a different set of technologies than we have seen in the past.

Matt Brady: 7:37 And I would argue, generally speaking, it’s very similar to other types of computer science and data-based technology. In a few ways, it is different. And I think we can be cognizant of those things. But I see a future with more jobs, more enjoyment, broader engagement, and hopefully better outcomes.

Claire Stewart: 7:58 I like the sound of that. Sounds great. So, looking at the broader landscape, leaders are under immense pressure to adopt AI rapidly. And that’s what we’re hearing day in and day out, is how it needs to be used more and used faster. So, how can executives embrace AI and automation at scale without losing the sight of empathy and ethics and human judgment that we’ve been talking about?

Matt Brady: 8:21 Well, jobs are changing, but I don’t think they’re disappearing. Students hear AI will take every job. I don’t believe it. I don’t believe it. I’m not seeing it. There have been shifts in the job market. But some of those shifts were due to the fact that during the pandemic, there was a defensive hiring strategy where large tech companies, especially, were gobbling up all of the talent. The interest rates were low. ZIRP, if you’re familiar with the acronym, zero interest rate policy. And so, when capital is effectively lent for free and talent is abundant, a company that has high hopes for its future is compelled to hire up very proactively. And I think there were significant corrections due to that. There are other financial policies, and let’s say policies domestically and abroad, that affected what we attribute the AI job, you know, doom loop to occur in ways that I think is misattributed in some cases.

Matt Brady: 9:20 All that is to say, be wary when fear is the sales pitch anywhere. I think that whether it’s a utopian view or a dystopian view, those, I think, are pretty safely categorized as bookends to common sense and likely outcomes, which hopefully the bell curve yields much more likely outcomes. Start with people, then pick a tool. In any hiring, in any equipping or enablement, I think it’s always good to start and end with the people. So, for our students, I think they have a bright future. I think there’s a tall order to step up and really determine how technology affects business. So, speaking from a leadership-centric point of view, I think all of our students should be technologists and trained appropriately in AI and other leading-edge technologies.

Matt Brady: 10:11 But to be clear, AI should augment people, not replace them in leading, mentoring, counseling, etc. And I think our students are the ones who will lead that change.

Claire Stewart: 10:23 Do you see that there are any companies or governments or civic organizations that are using AI and human-driven judgment really well together as they grow?

Matt Brady: 10:34 That’s a great question. One that comes to mind is a company I had some direct interactions with in the startup world. They’re a very large, well-known technology company. At the beginning of this year, they had 2,000 coding agents. And at present, they have scaled that to 20,000 coding agents. So, an order of magnitude of growth, working 24/7, never needing a break, doing good, if not great work. And I think the remarkable insight that I gained recently is that they project, by the end of next year, 200,000 coding agents.

Matt Brady: 11:11 Now, people will still solve the hardest problems and be responsible for overseeing the quality and quantity of the results that are produced. But there are certainly very drastic changes occurring by way of agentic AI, where these agents can be more autonomous, more capable of doing deeper and wider job functions. So, there will be ramifications from that. On a smaller scale, the organization that I recently launched, called WorksBetter, is a very small team of people who have been able to make sense of what, at this point, is 180 different public sources, all that are readable by humans and machines.

Matt Brady: 11:56 And what we’ve been able to do is address this question of, what is the state paying for near you, near a taxpayer. And how do tax dollars… what are the sources of tax dollars? How do they flow out of our wallets collectively? And then how do they get used? So, the uses of funds back to local communities, to the state, to other broad initiatives. And that would not have been possible without agentic capabilities, which are well known from many of the frontier companies and labs, as it were.

Matt Brady: 12:33 Whether they’re paid or open source, these tools are equipping small teams of entrepreneurs with big ideas to really go after big value creation. And I think we’re just seeing the start of that model. From a skills perspective, what comes to mind for me is, and this is something that I really try to instill in my students, largely from my own personal experience, leaders are going to have to be comfortable not knowing. Not knowing how to solve a problem, not having sufficient data, not having experience in a particular domain.

Matt Brady: 13:11 With that confidence or that comfort comes great potential. Because equipped with this data-driven decision-making and these agentic capabilities of AI, we can keep judgment and ethics in place as a tool over the knowledge that we can gain. So, as many professors and teachers in general have spoken to their students and colleagues for a long time before me, lifelong learning is very much the way I see our industry and our society moving forward. So, I think there are remarkable opportunities for individuals, teams, and organizations to do what they haven’t done, to know what they haven’t known, and to go where they didn’t know they could go.

Matt Brady: 13:59 So, I, again, derive great hope from that potential.

Claire Stewart: 14:03 Yeah. I love what you said about, like, be comfortable in the not knowing, and not always knowing in this moment, kind of what’s coming, and then what you have the capabilities to do. I love both of the examples of the small and the large industries that are using this. So, from small businesses all the way to CEOs and executives of large businesses, is there some skill or mindset that you think they need to be successful in the current technology landscape, aside from, like, that comfortable not knowing? Is there something else that you would suggest for leaders kind of trying to pivot this space?

Matt Brady: 14:33 Yeah. I think it is incumbent upon all of us, as individual contributors, as managers and executives and members of our society, to think about the value that we can create and to leverage something like comparative advantage to do what you do best and trade for the rest. Creativity, problem solving, critical thinking, collaboration, consensus building, I think these are all skills that are going to become more important as certain functions and skills are commoditized. And I think, again, that argues for a human in the loop or the human in charge of these technologies.

Matt Brady: 15:18 But some of the supporting skills therein would be delegation, the identification of business processes, simple optimization. I’ve always taught that you never want to automate a bad or ineffective process; that will just make it happen faster. So, automating processes that have been carefully thought through, improved on paper, simplified, and then optimized before they become automated is a first principles approach. And it’s one that I think actually argues for common sense, but I see students, technology, employees, educators, policymakers, really recognizing an opportunity to become the boss of the work that they used to do individually and to delegate out and to decide what “good enough” looks like and what “done” is defined as.

Matt Brady: 16:14 And then farm out or ship those requests and requirements, wait for the results, whether it’s seconds or much, much longer, that they have to wait to have the work product that they expect. But I think that we can all become producers of greater value by running “Us Inc.,” “Me Inc.,” and “You Inc.,” in producing what may have taken longer and not been done at the scope and scale that we can now achieve. So, it’s a skill set that will require everyone to think about the results they deliver, the value they create, versus the “work” they do, which will certainly shift.

Claire Stewart: 16:49 That’s awesome. So, the relationship between ethics and innovation feels more essential now than ever. And I know that you’ve done a lot of work to develop curriculum here at Leeds in that realm. Tell us about what that looks like and why it’s so important for students to understand this at this point in time.

Matt Brady: 17:05 If innovation pushes the boundaries, we owe people a look at the second-order effects of what the impact is. I want to give a lot of credit on this topic to the folks who raised me and mentored me, and that continues in some cases to this day. But specific to my work at Leeds and CU Boulder, the Daniels Fund has been very supportive of ethics and decision-making, and they allowed me to be an ambassador through a fellowship program that they have where I developed some content called DECIDE BETTER, which I’ve delivered to hundreds of students and actually many hundreds of executives as well.

Matt Brady: 17:44 So, the three courses that I teach are all very focused on innovation. The first is AI and mobility, second is CRM, and the third is citizen development. These courses and my other work at CU, I roll into a program called HumanCentered Technology. And the through line on all of these courses and programs is ethics and empathy built into innovation as something that, if I were to try to describe my professional pursuits, that would ultimately be the common thread. How translation now lets us reach 8 billion people, that’s a new responsibility as well.

Matt Brady: 18:22 We can communicate with anyone in the world, any time of day, regardless of culture, context, background, and in some cases, even access to technology. We can bridge and span so many different gaps and voids. That is, it’s really remarkable how connected we are. And so, I think there’s a need to address that notion of being able to impact a global audience in a way that deserves some care and some attention. So, in class, students use an AI teaching assistant that I’ve developed. It has guardrails. It’s grounded.

Matt Brady: 18:57 It’s, in some cases, extremely limited and stubborn. Because if you ask the AI TA earlier in the semester for help on the final project, it will, like a proper TA, be fairly measured and guarded. And later in the semester, as things are getting down to the wire, that AI TA will, of course, be a little bit more forthcoming with some assistance, as would be appropriate for any coaching or mentoring engagement. But the point is that it drives them back to the material, and instead of doing the work for them, it tries to create those learning loops and reinforces where they are falling short of their proficiency, their understanding of the content, to allow them to self-feed.

Matt Brady: 19:41 And that is, ultimately, my goal is to teach students how to become resourceful, to have grit, to have that appetite that is ferocious for learning and development. And so, the AI TA has been trained the same way, to guide them appropriately. Ultimately, we’re equipping the whole person, not just the technologist. And so, my goal is to be thoughtful about not just having a command of the buzzwords and the jargon and to be able to talk about the latest technology or shiny object, but to think about the implications of these technologies, to take issue with some of the ways that these technologies have been rolled out.

Matt Brady: 20:23 And even to choose, in some cases, to opt out of certain technologies that may not be relevant, that might play into a weakness or an area where they philosophically disagree. So, that’s totally acceptable. I, for one, love hearing students talk about gaming because I know that it excites them. And that is important to me, that they find personal joy and reward. I have never finished a video game in my life. And so, that is comfortable for me to enjoy, at arm’s length, something that is not an interest of mine, but I love that it is an interest of other people.

Matt Brady: 20:59 And the same could be said about my hobbies and curiosities, that it doesn’t have to be everyone’s cup of tea or what gets them fired up for it to be valuable and to be worth pursuing.

Claire Stewart: 21:12 Love that analogy. So, earlier this year, you taught a course that culminated in students traveling to Japan to conduct research and develop AI and automation solutions to address real-world society issues. Tell us more about what that course looked like and why you think it is valuable for our students.

Matt Brady: 21:28 I’ll give you an anecdote to start with. We met in Japan a 104-year-old World War II veteran who shook every hand of all of us that were visiting. He drew a picture for us to take home and we communicated with this gentleman, and he was truly a gentleman, so significantly that we shared so few words with him. Much of the communication was nonverbal. I promise that every student and instructor who was on that trip will remember that moment. For me, personally, my grandfather fought in World War II.

Matt Brady: 22:05 His flag sits behind me on my table. To speak with this gentleman who obviously served on the other side of that war was really a remarkable experience. I have a sense that my grandfather and he would have gotten on very well together, so it was a real honor to have that experience. More broadly, spending two weeks in Japan was a transformative experience for every one of us. It humbled each of us to know that where we’re looking at AI, robotics, automation, voice-to-text and text-to-voice technology, that these are real people who are involved.

Matt Brady: 22:43 And we have to put them at the highest level of priority when we think about the interaction model. So the course is loosely called Build in Japan, and that’s because we partnered with Nagoya University. We have 20 students here from CU Boulder, 20 students there. We’re in Tokyo and Nagoya for 15 days. And during that time, we conduct field research in hospitals and elder care facilities before any products are developed. So the formal title of the course is AI and Automation, or AI and Mobility: Technology for Tomorrow’s Societies.

Matt Brady: 23:19 And we talk about tomorrow’s societies because Japan and other Eastern and Western cultures are experiencing a decline in population or what is in some cases called population collapse, where there are not enough babies being born. There is a long lifespan and a wellness and healthcare function that allows for hopefully a very high quality of life well into a hundred years and beyond. And so this is a very sensitive topic and one that is not limited solely to Japan, one that will instead face many of our countries and societies globally over time.

Matt Brady: 24:01 So there are very, very specific and severe implications that we work through together, which is a pretty heavy topic for an undergraduate business course that is decidedly technology focused. So again, I would say here that what we’re trying to achieve is a fidelity and an alignment between the use of technology and the impacts and benefits to people and societies. One quick example, Anshin Care, Anshin being a Japanese term, is a medical adherence system for older adults. It took first place in our business plan competition and pitch fest that we hosted in Nagoya at the end of the trip.

Matt Brady: 24:43 And we were honored to have distinguished guests and judges from the university that we partnered with in Nagoya, as well as from industry. And they took first place because they built a smart pill dispenser that included an app for the caretakers so that there were basic communications between the recipient of the medication and the caretaker, such that there was a sense of a pulse check, of sorts. How is the person doing? Are they adhering to their medication schedule? But are they feeling okay? Are they happy or sad, maybe lonely on a certain day?

Matt Brady: 25:23 And there can be basic nudges that can go in both directions to advise the caretaker, but also to let the elder care patient know that maybe they should take a walk or be in touch with a friend that day if there’s distance between the two parties. It’s a remarkable project. All of the student projects were so far beyond what I think they were capable of. And I think speaking for the judges and certainly for the folks, myself and my co-director who had designed this program for about a year in advance, we were all amazed at what they could accomplish in a short period of time, thanks to technology, but much more so by being very empathetic in the way that they approach solving problems for people who they may not feel they have anything or much in common with, but that they really found a lot of shared humanity and joy in leveraging business and technology capabilities to bring value and to deliver attention and love to an audience globally.

Claire Stewart: 26:25 That’s great. What an amazing experience for the students. I mean, it’s amazing. You got to participate as well, but what an amazing time for the students. It’s so cool. I love the examples.

Matt Brady: 26:34 Thank you. I was honored. It’s a very humbling experience to be on the learning end of a teaching experience like that. But I felt like I was very much in the shoes of the students as to taking in so much. Overall, most AI courses, I think, begin with tools. This course begins with people for reasons that I shared with you and is, again, very much my trademark in my approach to technology and business.

Claire Stewart: 26:59 Very cool. So how have you seen students’ interest or approach to empathy and innovation change with the rise of AI in the last few years?

Matt Brady: 27:08 It’s a really interesting question. I think that students started to use ChatGPT 3.5 when it came out in November of 2022, and they did so in ways that were very experimental. I think they wanted to see what it could do, and so they just gave it a shot. I see students now have a much more diverse set of experiences or perspectives on the technology. Some are leveraging agentic AI in the ways I described, where their computers are running 24/7. They have techniques of dropping a little file folder in between their laptops so it never closes in between classes.

Matt Brady: 27:47 And I see other students who are leveraging AI for more difficult projects, for research, for growth, let’s say, in ways that are more reactive and specific in nature. And then I see students who prefer not to be terribly exposed to it and would rather either do things themselves, they would prefer to be more independent versus dependent on technology, which I highly respect, or who haven’t quite figured out how to leverage the technology effectively, efficiently, in an affordable manner, because there are increasing costs to leveraging these tools specifically.

Matt Brady: 28:26 And I would foresee these costs actually going up, though arguably they could go down as well with open source. So let’s just say price is not static. So there are a variety of viewpoints that I see students taking, which is to say they’re more informed now and have the ability to have informed consent or to make choices for themselves that make sense based on what their needs are, what their career aspirations are, their preference levels. I’m bullish on technology. I’m a fan of using technology and business process to go faster, cheaper, better, but not everyone shares that view.

Matt Brady: 29:04 And I think it’s important, especially at the undergraduate level, to be exposed to those different viewpoints, to have the opportunity to try different approaches, whether it’s a no technology type of approach or a fully enabled technology-centric type of view, so that students can find both their preference and when to apply different approaches. Because I think an all-of-the-above mindset can ultimately be very effective as well.

Claire Stewart: 29:29 So looking to the future, it seems inevitable that AI will continue to change how people make decisions. So if you could give our audience of business leaders one tangible rule of thumb to ensure technology enhances human judgment rather than replacing it, what would that piece of wisdom be?

Matt Brady: 29:46 Set the bar before you see the data and never make a decision that you can’t explain. Let me break that down. Set the bar before you see the data, to me, demands that you have to have your own perspective, your own point of view. Before you open the report or run the model, make note of the result that would change your mind. Say it to your team. “If the number comes in below this line, we walk away.” If you wait until after you’ve seen the data, it quietly moves the goalposts.

Matt Brady: 30:17 And you’ll always find a story that fits what you already wanted to do. That is called confirmation bias. The second piece of this then, never make a decision you can’t explain. This is absolutely vital that we convey this message broadly, or we will all fall victim to this. Here’s an example. The systems that we’re working with are very persuasive. They’ll hand you an answer with total confidence. So the test is simple. Can you explain how you got there in plain language to a board member, to a customer, to a loved one?

Matt Brady: 30:50 If you can’t say how you got there, the tool decided, not you. And I think that’s a test that we have to be able to pass with flying colors. Ultimately, the gut sets the bar, the data tests it, and you stay accountable for the call. So we have to train the hands, the head, and the heart. That’s my mantra. I believe so strongly in covering all aspects of a student’s educational journey through our technology courses, through Leeds. I think CU does a great job of this as a whole.

Matt Brady: 31:23 And so that’s why I’m so well aligned with the curriculum and the program.

Claire Stewart: 31:27 I love that answer. I need to clip that and keep it and put it up on my wall in my office, Matt. It’s wonderful. I really like that. Well, thank you so much for joining us today, Matt. It has been really insightful. And we have loved learning about your impact at Leeds, as well as your vision for the future of business and how it works with technology. So thank you so much for being here.

Matt Brady: 31:46 Thanks, Claire. It’s an honor.

Claire Stewart: 31:48 Thank you again for listening to Leeds Business Insights. Make sure you’re one of the first to hear every episode by subscribing to the show wherever you get your podcasts. The Leeds Business Insights podcast is a production of the Leeds School of Business and produced by University FM. We’ll see you next time.