Bill Kerr: Is trust in robo-investment advice misplaced? It may be too soon to tell, but the acceptance of automated guidance helps explain why investment services companies are among the most AI-exposed. At the same time, investment research firms are key sources of analysis on the effects that automation, climate change, diversity, and other factors are having throughout the market. And, of course, they’re turning to AI to aid in that analysis. What does this mean for the sector’s workforce, for the competition for skills, and the relationships between human experts and their customers?

Welcome to the Managing the Future of Work podcast from Harvard Business School. I’m your host, Bill Kerr. I’m joined by Kunal Kapoor, CEO of investment services firm Morningstar. We’ll talk about the sweeping changes redefining the industry and the skill set professionals need to provide investment data, analysis, and advice. We’ll also talk about the social and political pressures influencing perceptions of sustainability and diversity as business metrics. And we’ll consider how workforce strategy stacks up as a variable in market ratings and risk analysis. Kunal, welcome to the podcast.

Kunal Kapoor: Thanks for having me, Bill. Happy to be here.

Kerr: Kunal, unlike many career trajectories today, you’ve been with Morningstar throughout most of your career. So tell us a little bit about that experience and what helped you get started and work up the ranks?

Kapoor: That’s right. I started in Morningstar a week after I graduated college, and so I’m often greeted with the question right up front as to what I’ve been doing at one firm for more than 25 years. The reality is, I feel like the firm, itself, has had many avatars over the time I’ve been here. The job that I first did when I came to Morningstar doesn’t exist anymore, and the businesses that make up Morningstar today are also in some ways not all ones that existed when I started. We were one office here in Chicago, with about 200, 250 people when I started, $30 million in revenues. And today, we’re a global firm with 10,000-plus colleagues around the world, and we’re serving customers across the investment spectrum in different business lines. And so the firm has had a bit of a metamorphosis over time. And as it has shape-shifted, it has provided lots of opportunities. That being said, the one thing that is important that you hang on to, no matter what, is the underlying set of values that define a firm—because I do think those are super important—and the mission. It’s why you get up and come to work and what energizes you to feel like you want to contribute and make a difference.

Kerr: And has that mission changed over the couple of decades that you’ve been there, or has it stayed consistent?

Kapoor: Ultimately, the wording of the mission has evolved a little bit. When I became CEO, I refined it a bit further, and it’s now all about empowering investor success. But that concept of succeeding when the investor succeeds and doing the right thing for the investor—because ultimately, if the investor wins, so does Morningstar—has underpinned our mission in one form or the other in the time that I’ve been here.

Kerr: So go back to the job that you first took on no longer exists, and maybe give us a high-level understanding of how the role of the financial adviser has been changing with all the technological advances over the last couple of decades. And how did the company look, and what does it look like today?

Kapoor: Let me start first of all just with Morningstar. And a simple way that I would describe it is, when I started 25-plus years ago, the portfolio was fairly simple for most of us. It was stocks, bonds, and cash, in one format or another, and that’s what most people owned. And, candidly, there probably was a little bit of a tilt in those days to safer fixed-income instruments. And so you saw a lot more cash and bonds than perhaps you do today. Over time, however, that portfolio has changed. And so today, we’re talking about private equity assets, venture cap assets. A fixed-income mutual fund, which I remember would just own bonds, today gets its exposures through futures and swaps and instruments that don’t directly own the bonds but get the right kinds of exposures that folks are looking for. And then you have had all kinds of other types of vehicles come along, like ETFs; you’ve had the rise of what are called “model portfolios,” “separate accounts,” and “personalization.” And so the change for Morningstar has been that we’ve gone from collecting just very narrow sets of data to much broader sets of data, because our goal ultimately is to X-ray, as I like to say, the investor portfolio. And so, as the investor portfolio changes and evolves, we’ve needed to as well, so we can X-ray that and explain to an investor what’s going on in that portfolio and whether it’s going to help the investor reach her or his objectives. So that’s always the intent. And then the financial adviser has changed along the way as well. But the most important thing that I’d say is, if you ask someone of my age—in their 40s, 50s—what the typical picture of a financial adviser is, you just get this answer often of a person who walks into someone’s home with a little bit of a briefcase with lots of papers and sits down and spreads it on the dining room table. And you have this conversation off of those papers. And today, that couldn’t be further from the truth. It’s so much more automated. And advisers have learned to scale their businesses and scale what they’re doing. And so the average size of the client that the adviser serves is much larger. The types of services that an adviser is able to bring are quite different. And I think, more importantly, the adviser has become a partner and coach, almost, to their clients in a way that perhaps even goes beyond their financial lives. And so the adviser has had to rethink the way he or she operates, and technology has been a big enabler of that.

Kerr: So continuing on the Morningstar side, tell us a little bit about the portfolio of talent and skills that you have inside the company, and what’s the typical way that you’re recruiting and training those individuals.

Kapoor: We like to say that we can always teach finance, but we can’t teach you how to think. And so that means we don’t get too hung up, especially at the undergrad level, if you’re walking in without a finance degree. We certainly like folks who have finance degrees, but we’re really open to those who have shown proficiency and a broader aptitude. And so we hire as much for what we perceive as intellectual curiosity, hustle, energy, perseverance, those types of skill sets, as much as anything. And this includes folks on the technical side. You certainly want people entering the technical side to have a set of technical skills, but we’re looking for people who can do more than just have technical skills. And candidly, even on the technical side, what you learn when you’re learning programming in school and when you walk in and have to do it at a product level, or you’re programming a database, it’s just a different experience. And so the on-the-job training is incredibly important. And our ability to hire people who are versatile is pretty critical in that context.

Kerr: So, not surprisingly, we’re going to want to talk about artificial intelligence in a variety of ways. But let me start with your overall take about AI and its strongest use cases.

Kapoor: From my perspective, the most exciting near-term use cases are not that complicated, and they’re really geared to taking friction away from all of us in the way that we operate. So think about it from the perspective of a Morningstar client, for example. And the financial adviser that you cited is a perfect one that we could focus on. So a financial adviser who’s been reading Morningstar’s research, looking at our data, up until recently, you’re probably going to run some screeners, you’re probably going to run some searches, you’re probably going to look at old content, you’re probably going to read some analysts who you trust that cover the sector. But you’re going to spend some time finding what you need. It’s vastly improved from what it used to be, but there’s some friction there. Now, you fast-forward to what we’ve done, and so we took all our data and research going back to the founding of Morningstar, fed it into a database, and created our own chatbot on top of it using OpenAI’s technology and put an avatar on top of it. And today, a financial adviser can just talk to “Mo,” as we call the avatar, and ask Mo, “Hey, Mo. How do I know what Morningstar thinks about investing in AI?” Or, “Hey, Mo. What are Morningstar’s best ideas for investing in AI technology?” And boom, you get an answer just like that. And you suddenly think about the time you saved and the friction you’ve removed by not having to do the search or do the screen or figure out exactly what the parameters are. And that’s pretty powerful. And it’s a simple case of reducing friction. Or you’re in our products, and you may have previously gone to the FAQ and tried to find the exact answer. And we’ve all lived this—like, “Oh, where’s my question? I need to find that one answer.” And you’re like, “Oh, there’s so many questions here, but I can’t get to that one.” You don’t need to do that anymore. You just say, “Hey, Mo. Help me find X,” and boom, you’re there. And that is taking friction away. And that is not, by any means, a complicated use case. There certainly are many complicated use cases, but I like those examples, because they can be applied to so many facets of our lives. And they show that so many things we do today that we may think are time efficient are actually not, and we can make them even more efficient and more impactful in terms of how we use them. And I’m really excited, at least in the near term, about AI’s potential in that realm, to reduce friction in that sense.

Kerr: Many of us will have used ChatGPT in various ways. And, of course, we all know that you can have two different people put the same prompt in and the responses will differ, unlike the old static frequently-asked-questions page, where everyone gets the same four-sentence answer to something about Morningstar. In Mo’s case, does it provide different answers for individuals, even if they start with the same initial set of prompts?

Kapoor: We’ve tried to get around that problem by ensuring that Mo cannot leave our architecture. And so Mo is in Morningstar’s walled garden. And if you’re going to ask about AI ideas, Mo is using the same set of data and research to answer your question. So, with the example that you cited, the reason we all get different answers is because there’s not a walled garden. And so we’ve restricted the ability of Mo to exit the garden, and that provides for greater consistency.

Kerr: Sometimes you hear about financial firms—or investor PE companies and so forth—using the generative technology, where you can get different answers to try to get a range of outcomes. So they want to see the variation that they get. Is that something that you see a power to or a benefit, or is it more of an anomaly that probably shouldn’t be baked into calculations?

Kapoor: Well, I think it depends on what type of investor you are. My guess is that some of the investors you’re citing tend to be macro-oriented investors, who are trying to gain their edge by sorting through reams and reams of data, and they’re already in the business of doing that. And what you cited is probably just another tool they can use to gather even more intelligence in that sense.

Kerr: This came with your “walled garden” approach to this, but thinking a little bit about the errors that generative AI can have, are there safeguards that you put into place around its role for market analysis or financial advising or things that you would advise your clients to put into place?

Kapoor: Our biggest safeguard right now is that we don’t let it leave that walled garden. And so we know exactly what it’s being fed and what it’s drawing from. And so the data and research that it’s drawing from have already gone through editing cycles, have been quality checked. And so we’re very confident in it, because it’s already in the hands of our clients in that sense. And so we don’t allow anything else into it. And that has been the way we’ve approached it. And I think having private instances is a helpful thing. And you can extend it to internal employee use cases as well.

Kerr: Tell us about how you are—at Morningstar, just in general—building your AI strategy. Is it something that’s very centralized? With 10,000 colleagues around the world, is it decentralized to business lines or units? What’s the approach you’re taking toward figuring out where the company should be investing and how?

Kapoor: Yeah, that’s a really good question, and I think I’ll start by saying that we acknowledge and feel that there’s a need for speed here. And so, when you’re looking for speed, inevitably you want to have some level of decentralization. And at a high level we like that. But because the technology’s new—because we want to be sure that any work we do is secure, it’s ethical—what we’ve done is, we’ve formed an AI council. And the AI council consists of senior folks from our technology department, senior product leaders, as well as a couple of key leaders from our legal department. And they have essentially set the policy and reviewed use cases, when they come alive, before our teams are able to go out and experiment with those use cases. So what we’ve tried to do is, we’ve tried to centralize the vetting of ideas and the use cases, as well as how they’re going to be done. But we’ve tried to decentralize a little bit the teams that are executing against it, although we are building a central set of tools that we want everybody to use, including what we call our “intelligence engine.” And the idea is, if we have these sets of tools, even those who need to build on them will be able to move a little bit faster, versus doing all the work themselves.

Kerr: Stay a little bit longer on that, because you highlighted how you’re selecting where to try things out, to experiment, to build the early applications and use cases. One thing that many companies struggle with is, “Then we have these smaller things that we’ve created. We want to scale them up, we want to take them across the organization.” How have you been orchestrating that when you see something that’s working in one part of the business ported to other aspects?

Kapoor: Well, the way we try to do it is to, first of all, ensure that, if we’re going to require something to be used across the business, that it’s truly something that can be leveraged and is useful across the business. And so Mo is a perfect example, in the sense that we fed it all the research and technology. And on top of it, we can build search scenarios, we can think about service. Service is a great example. We started to use Mo in the service context. And so, if folks are calling in, even our internal phone reps, for example, can use Mo to figure out how to get the answers quickly. And so you think about that, and we can take that and scale it across every business within Morningstar, because the core technology there that we’re using for Mo is similar. However, if you move from one part of the business to the other, what you’re changing is the user manuals, the methodologies behind certain products. And so you’re feeding all that into the technology and then getting answers appropriate to that part of the business. So I think those are some ways we can do it. Having said that, within some of our products, we allow a little bit more experimentation, because the use cases start to then be different in terms of how a user will come in and interact with the product—and especially if you’re thinking about friction and removing friction. A financial adviser may be doing one thing, a PE firm may be doing something entirely different. And so, within products, then, we have to be sensitive and think about product use cases quite differently.

Kerr: Kunal, you earlier mentioned a talent strategy that relied on bringing people in that would think in creative ways and you could teach them the finance side.

Kapoor: Think critically is what we always think about.

Kerr: Yeah, critical. How are you approaching the reskilling toward AI? Are there specific things you’ve done to help your employees go up the learning curve?

Kapoor: Absolutely. And so we started with the technology teams, and as you’d imagine, we’re making different technologies available to them. We’ve been reskilling them through various coursework, mostly online coursework. And then, the most important thing that we’re also pushing—and I think there’s no substitute for this—is actual experimentation, and the experimentation piece, in particular, we’re trying to push very broadly. And so one of the most important things we can do and that we’re trying to do is trying to think about ensuring that people are using and thinking about this technology regardless of the work that they do. And so an area, for example, with lots of data analysts across Morningstar—because we’re covering so much when it comes to securities data, when it comes to company data, when it comes to ESG data—so a lot of these folks collect some of that data manually, particularly when it’s nonstandard. And that’s a perfect use case, where we’re asking them to experiment with the technology because we think much of what they do that is laborious can start to be moved to a more automated process thanks to AI. But they’re the ones who are going to help us solve that, as opposed to someone who doesn’t interact with the data, who may be a few steps removed from it.

Kerr: Yeah. So it goes back to your earlier conversation also about the centralized screening and so forth. There’s this fascinating blend of both the organization pushing in some directions but creating the micro experiments and awareness throughout the company.

Kapoor: Exactly. Exactly. Technology and using technology is a great way to differentiate and propel your career. It’s just a reality. And no matter the role you’re in, if you think about it, you can do it. And I think there’s a general belief that programming is something that you have to have very strong training on, or you can’t do it. But the reality is, the tools today, even with the copilots, essentially mean that someone with very minimal training up to this point can essentially over a weekend figure it out and, with the copilot, do some level of programming. And that’s pretty powerful. And if you think about reskilling and how you can do it yourself just by working at it a little bit in your own time, as well.

Kerr: We spent a lot of time talking about AI inside of Morningstar. Tell us a little bit about how it’s impacting the ratings of businesses, whether it’s their exposure to it or their proficiency in working with the new technology.

Kapoor: Yeah, great question. So we recently crafted a new index. It’s called the “Morningstar Next Generation Artificial Intelligence Index.” Snappy name. And what the index does is, it provides exposure to companies that are producers and suppliers of generative AI, and which we then expect to earn a significant amount of revenue—"material revenue,” as we call it—and net profit growth from generative AI. On the downside though, our equity analysts also are thinking about assessing the risks of disruption—from technology or otherwise—and obsolescence, when thinking about our coverage lists. So AI is a newer way for businesses to potentially be attacked, but companies with economic moats, we think, should be able to fend off such attacks. And on the upside, we think—and this is our analysts’ view—that certain companies and industries are going to be more efficient and gain new insights by deploying AI, although it’s fair to say that we’re trying to quantify those insights on a case-by-case basis, and it’s not super easy to do that quite yet today. Across our tech stock coverage—given that that sector, in particular, has a lot of AI exposure—the primary developments that we’re paying attention to include cloud computing, designing AI processors, and also building foundational AI models like ChatGPT. However, I just say that there’s many more companies that profit from these trends and support the Gold Rush. People sometimes forget that the companies that provide the tools to do the digging can be just as interesting to invest in. And so we look at data and infrastructure, software, IT services, and we look at companies with healthy revenue and profitability exposure that are, thus, likely to be considered in this AI index that we’ve built. The landscape’s changing rapidly, but it’s clear that the early winners so far are ones that are not exposed to a single model. Instead, it’s companies like Nvidia, which has obviously been meaningfully in the news, which is basically serving as that tool supplier to the businesses and cloud computing firms that are benefiting from AI and deploying all these models. So you want to be thinking about it in that fashion.

Kerr: Hardly a conversation also goes by without having some reference to environmental social governance, ESG-type goals, or DEI, diversity, equity, inclusion-type work. And that’s something that you’ve been researching and has a part of the product perspective around. It also, of course, interacts with politics today. So tell us a little bit about the state of play of those efforts.

Kapoor: I’m going to steer clear of the political aspect of it, because it’s not only tricky to comment on, but the reality is, it’s different in different parts of the world. Acceptance of ESG in Europe is incredibly high, and table stakes in the U.S, I think, we’re trying to find our footing in terms of what it means. And sometimes it can get boiled down to topics that are very politically divisive, such as fossil fuels, for example, and whether you use them or not. But in reality, there are many other use cases that are far more interesting, in my view, such as: Do you want to build a portfolio that supports local businesses? Do you want to build a portfolio that supports, perhaps, your religious beliefs? All kinds of interesting things that you can tilt toward. And personalization is, I think, going to emerge over time as a very strong tool—whether it’s through AI or other means. Today, you and I, Bill, if we want to invest in certain things, we probably own investments that are similar on many fronts and that don’t personalize to us, and that has been the case because it’s been expensive to do that. That will no longer be the case in the future, and it’s our job to collect all the data that allow people to make the choices they want and to express their preferences in whatever way they may choose to. That’s what makes markets. You don’t have to agree. And it’s not a right or a wrong. It’s a preference, and you live with the consequences. And it’s perfectly reasonable for people to use the same set of data and come to different conclusions.

Kerr: A thing that’s obviously very important in the Managing the Future of Work context is the talent base of a company, its workforce. And as we enter environments where there’s ever less labor abundance, more labor scarcity, and as you get into questions of how good is a company’s workforce toward things like the future of AI and generative AI and so forth, those could be very material for an investment decision. So are you able to capture some of that stuff that’s inside of organizations and make that part of your ratings and risk analysis?

Kapoor: Yeah. It’s a slightly tricky topic, and we have something that we call the “human capital material ESG issue,” and that’s a building block for our aggregate ESG risk rating. And what we do is, we assess a company’s human capital risk based on two factors. The first is how vulnerable they are to experience human capital risks based on the characteristics of their workforce, and secondly, how well a company manages this risk through policies and programs, such as adequate hiring and retention policies, employee turnover, employee development, things like pay transparency, workforce diversity, KPIs, mentorship programs. And so we look at all those types of things. And human capital is one of the most frequently material ESG issues for companies assessed in the context of this risk rating, and it reflects, I think, the importance of the workforce in creating and driving economic value. And so that’s how we’ve tried to approach it.

Kerr: As you think about the next five years, the stuff that lies ahead—and we’ve covered lots of different topics along the way in this podcast—but what are some of the things that are additional key trends you see playing out that you want to be both advising on or having the data to support the financial advisers with, and also that could impact Morningstar, itself, the organization?

Kapoor: I’ll end where I started, which is, I think portfolios are going to have more options in them than ever before. I think they’re going to be more personalized than ever before. And I think they’re going to be cheaper than ever before. And so, for us, that means we need to ensure that our databases continue to grow and are ever-expanding, so that we can cover what’s going in those portfolios and help people personalize, because I do think that’s going to be a very important factor in the future. One of the things that exists today is that, more and more, in some parts of the world, including here in the U.S., individuals are responsible for their retirement finances. And one of the ways we empower them is by allowing them to engage with their finances by personalizing the data. And I think this point is sometimes lost on those who just look at investing from a “did you underperform or outperform” type of perspective? Because the reality is, most people are not engaged with their money, and they either find it hard or not interesting to be engaged. And so one of the societal risks is that, as they approach retirement, they’re not ready to retire from a financial perspective. And one of the most important ways that we can live our mission is ensuring that we have the data to allow personalization, which I strongly believe leads to engagement and the likelihood that people are going to save more and engage with their portfolios and, ultimately, end up with better outcomes. And so that’s how I think about the business.

Kerr: Kunal Kapoor is the CEO of Morningstar. Kunal, thank you so much for joining us today.

Kapoor: Thanks for having me. I enjoyed it.

Kerr: We hope you enjoy the Managing the Future of Work podcast. If you haven’t already, please subscribe and rate the show wherever you get your podcasts. You can find out more about the Managing the Future of Work Project at our website hbs.edu/managingthefutureofwork. While you’re there, sign up for our newsletter.