Forty percent of Black American households don’t have high-speed, fixed broadband. That digital divide is just one of eight investment opportunities that, according to recent McKinsey research, have the highest potential to move the needle on Black economic mobility. The others are affordable housing, pre-K–12 education, health equity, financial inclusion, business ownership, workforce training and job attainment, and public infrastructure. On today’s episode of The McKinsey Podcast, McKinsey senior partner Shelley Stewart III speaks with editorial director Roberta Fusaro about how impact investors can help close the digital divide, create more financial inclusion, and achieve equity in healthcare for Black Americans.
After, Asaf Somekh, the CEO and cofounder of Iguazio, a McKinsey company, talks about artificial intelligence and offers some solutions for addressing pilot fatigue.
The McKinsey Podcast is cohosted by Roberta Fusaro and Lucia Rahilly.
This transcript has been edited for clarity and length.
Financial inclusion occlusion
Roberta Fusaro: Hi, Shelley. Thanks for joining the podcast today.
Shelley Stewart: Thank you for having me. I’m very happy to be here.
Roberta Fusaro: Black economic mobility is why we’re here today. Our recent report identified eight areas with the best potential to improve it while also offering investors favorable market returns. We don’t have enough time to discuss all the areas, so today we’ll focus on three: financial inclusion, health equity, and the digital divide. But before we get into all that—with the rise of impact investing, would you say now is a particularly opportune time to reach out to investors?
Shelley Stewart: I think that’s right. What happened to the Black community and other underserved communities during the pandemic, as well as some of the social-justice movements in the wake of the murder of George Floyd and countless others, really brought this conversation of economic mobility into sharper focus for folks who may not have been spending as much time thinking about it. You see it proliferating throughout the economy, throughout sectors. You see it in the public sector, the social sector, and in the private sector, including the investment community.
Roberta Fusaro: I want to focus first on financial inclusion. There is an 80 percent wealth gap between Black and White Americans. What are the reasons for this disparity?
Shelley Stewart: Sometimes people assume that the lack of involvement in the financial-services sector is just a symptom of a lack of resourcing in the Black community. It’s important to think of it both as a symptom and as a cause. Many of the traditional barriers to Black Americans in the banking sector have led to a historical distrust, as a function of interactions over time, that includes being excluded or interactions that have been less than positive.
There is another point just around access, particularly in a kind of analog and brick-and-mortar world. There’s this notion of banking deserts and not being proximate to actual branch banking. This creates an actual physical hurdle around convenience and access. And the other things I would cite are the products and the offerings. By virtue of having smaller incomes and lower or no starting wealth, some of the products are just cumbersome and not appropriate with the form factor today.
So it’s just not working for swaths of the population. And that means that people are not saving. They’re not able to access their checks in ways that are low cost, like direct deposit and then you take your money directly out of your bank. You don’t have to go to a check-cashing place that might take a larger percentage of that to help you or bridge through a payday loan. That creates challenges, and that’s where the private markets come into play.
Financial impact investing
Roberta Fusaro: To that point, what can investors do? How can folks invest in a way that will make the greatest impact?
Shelley Stewart: I have a few thoughts about where the opportunities are. The first is about trust. The lack of trust in existing institutions creates a clear opportunity for new entrants, nonincumbents, to enter the arena and develop a compelling value proposition—and loyalty.
The second point is about access. There’s the brick-and-mortar point that I made, but as we think more about the role of digital, and the dramatic disruption that we’re seeing within the financial-services sector, that creates real investment opportunities for people to develop platforms across banking, credit products, and investing products that are easily accessible on someone’s device.
Want to subscribe to The McKinsey Podcast?
And that creates private-capital opportunities for investors across a number of different areas. And you’ve seen capital flowing, particularly in the start-up realm, around some of these, but also in some later-stage companies.
And then the last point is, it’s not just about access to products—it’s also about how digital and some new and innovative ways of approaching things, like credit scoring, create opportunities for new products and services that didn’t previously exist. How do we use this opportunity to develop offerings better suited for people with some of the income and wealth-starting constraints the Black population has?
Thinking out of the box
Roberta Fusaro: Are there examples of the progress we’re making in this area?
Shelley Stewart: Yes, I think there are. I think you’re seeing more and more digital banks going after this underbanked or unbanked segment of the population. And if done well, the operating model allows you to have a fee structure that works better for this group because you don’t have all of the fixed overhead and ongoing costs.
I think you’ve also seen on the investing side some apps that offer simple investing. I’m not talking about complex things. That’s a story for another day, because I think there are a lot of changes that are also going on that are not appropriate. We’re always mindful of that, particularly when you’re talking about financial services, that there’s this delicate balance between predatory and constructive.
But you’ve seen some stock investing in other apps that lower the barriers on whether you can buy one share or partial shares, and it doesn’t cost that much to transact. So I think there’s some innovation going on.
And then on the mortgage side, you’ve seen some innovative things from what I’ll call nonincumbents—and even from some of the incumbents—like requiring lower down payments and trying to get to this idea of, “How do I tailor the product so it’s fit for purpose?”
I think you’ll continue to see investment opportunities in and around that realm, whether it’s mortgage products or even developing innovative ways to better assess credit. And if you can do that, and if you can price credit transparently, well, that’s a business opportunity. And I think people recognize that, and you’re starting to see capital flow.
Not everyone has access to healthcare
Roberta Fusaro: Yes, the more success that you see in certain areas, with certain products, it sort of engenders more and more investment.
I want to switch to the health equity topic. Again, I’m looking at the report, looking at figures, and we see in those figures that roughly three million more Black Americans would be alive today if we had closed the disparities in healthcare between Black and White Americans. What are some of the reasons why these gaps exist?
Shelley Stewart: I think almost every discussion about economic mobility and disparity should start with health. If you don’t have health at the foundation, then you can’t really participate in the economy in a meaningful way.
If you don’t have health at the foundation, then you can’t really participate in the economy in a meaningful way.Shelley Stewart
There are a lot of systemic reasons for why we see some disparate outcomes for certain populations. I’ll just give a few thoughts on some of the challenges that surfaced in a lot of the work that we look at. The first will sound familiar, but it’s around access. And just like you have financial-service or banking deserts, you’ve got healthcare deserts.
So not being proximate to care is a challenge. People are living busy, busy lives. I think sometimes we conflate income with how hard people are working, and that’s an incorrect conflation. You have people who are working two, three, and four jobs. So if access to care is not proximate in the context of them working many jobs and taking care of their families, then it is a real challenge.
The second one is a spillover from the labor market challenges that many Black Americans have, whether it be the sectors that they’re employed in, or the types of jobs they have, and what that means in a market where much of the insurance is a function of your job. That’s the model that is largely in place. And if Black Americans are not participating the same way that other groups are in the workforce, and in jobs that actually confer these benefits, then they’re going to be underinsured.
And that has huge implications for the type of care you get. It’s a double whammy because the cost of healthcare when the need for it becomes acute—because you don’t have the access to preventative care and because you’re not insured—is higher.
If you combine the lack of access with the insurance coverage issues, and then you overlay what we know shows up as bias—you’ve seen some of the recent stories about the mortality rates of Black women who are pregnant compared to other segments of the population, even when you adjust for education level, and it’s staggering—and if you take those things all together, that presents a real challenge, but it also presents a real opportunity.
Addressing inequities in healthcare
Roberta Fusaro: I think there was something in the report, as well, that talked about how healthcare providers were prescribing pain medication differently based on their own racial bias. How can investors help move the needle in this regard?
Shelley Stewart: The investor community alone won’t solve this one, but that’s the spirit of this whole discussion. It’s going to take the private sector, private capital, social sector, and public sector coming together. I do think there are a few opportunities here. One of the big social determinants of health outcomes, way upstream, is food and nutrition. When people think about health, they don’t always go there, but this is a place where we should start.
We know that if you’re Black, you’re significantly more likely to live in a food desert, which is a USDA-designated census tract that does not have what is deemed to be adequate access to fresh food. So how do we think about investing in these communities in the food supply chain, if you will, to bring healthier and better alternatives to these communities? And by the way, we spent a bunch of time surveying thousands of Black consumers. Their preferences for healthy, organic foods are stronger than other consumers, on average.
The demand is there. So how do we find ways to invest in businesses and business models that can bring that healthy and nutritious food to these communities? That’s one idea. I think a second one, as we think about our healthcare system broadly, is how are we investing in ensuring that when you do things like develop molecules and drugs, or when you’re training physicians and caregivers, how do we make sure that we are injecting this idea of diversity of patients all throughout that journey? And I think that represents an opportunity for investors who say, “What are the business models, and training programs, and things that we can overlay inside of the system to help accomplish that?”
Even more extreme, if you think about life sciences companies that develop molecules and drugs, one opportunity for exploration that we’ve talked about is there are certain drugs and conditions that are more prevalent in Black populations.
That’s an opportunity to go nurture some drug or molecule that might be underinvested in if you took a more generic or broader lens. How are we finding these niche opportunities to capitalize things that will disproportionately help certain communities that have maybe been under-supported in the past?
That’s the second thing I would say. And the third thing is, we start to merge the digital concept with the health to get to digital health. And that helps to bridge some of the access gaps that we’ve been discussing, right?
Also, how do we invest in digital capabilities to get to people in a different way, to help bridge some of those legacy issues, and at the same time invest in brick and mortar in these communities, from a health-provider perspective?
You’ve seen some models pop up, some of these smaller-format, fully integrated internal medicine or specialist health options. And they’re starting to show up more and more, particularly in urban markets. I truly believe that’s an investment opportunity, and you’ve seen some of them be incredibly successful.
The digital divide
Roberta Fusaro: Shelley, you’re doing my work by leading us into this next section on the digital divide.
One of the areas where we see a significant lack of access is high-speed, fixed broadband. Looking at one of the statistics from the report, approximately 40 percent of Black American households, as opposed to 28 percent of White American households, don’t have high-speed, fixed broadband. What’s preventing Black Americans from gaining this access?
Shelley Stewart: I think about broadband infrastructure in general as an enabler to address a lot of the other challenges, and I think there’s been a sharp focus on this recently. The administration in Washington has talked a lot about this issue. There are a number of noted philanthropists who have identified this as a core issue and a passion of theirs to address.
At the core, there are a few challenges, and they differ by type of geography. One is just the physical infrastructure itself. Also, you need real capital in the ground if you’re going to unlock it for folks. There’s capital showing up with some of the recent bills that have been passed. And how do we make sure that capital gets to the places where it’s needed most? And are the services affordable? Can folks consume those services in a way that is not overly burdensome relative to rent and the basic necessities of living? This is an opportunity but also another cost bucket.
The third piece is, once you’ve got [the services], then what device do you need to consume the whatever it is you’re trying to do, whether it unlocks not just financial services and healthcare but also educational opportunities? Do you have a device, a tablet, a computer that really allows you to engage beyond just your iPhone? And by the way, phones are great, but that form factor is not necessarily always conducive to all of the great use cases that we have to leverage digital as a real enabler for education, for health, and for financial services. So we also have to address the devices issue that sometimes gets lost in this discussion.
Where should impact investors start?
Roberta Fusaro: This report sets out a road map for investors and says, “Here are the eight highest-potential areas.” How can investors prioritize?
Shelley Stewart: We don’t have all the answers, and by no means is the research piece meant to be an exhaustive guide. It’s meant to be a starting point for where we have identified existing sustainable business models and marry that to where we’ve identified interventions that we know will improve outcomes for people.
So those are the areas we identify as investors start to think about, “How do we operationalize that?” What are the issues and areas that you’re most passionate about driving different outcomes? And then, how does that lead you down the road to, say, “I’m going to focus on the digital divide, or I’m going to focus on affordable housing, or I’m going to focus on banking.”
I think it really is trying to marry that view with your core competencies and your professional networks, or even maybe where you think your limited partners, the folks that are providing the capital to you, where they may have interests. Part of this is also to skate to where the puck is going.
So what our aspiration is in putting this out is that we will learn from the field. We hope that there’ll be five, ten, 15 more business models that emerge in each of these eight areas that are sustainable and that do good along the way. That’s really the hope.
Roberta Fusaro: Shelley, this has been a great discussion. And we’ve only just touched the surface. So let me remind listeners that they can visit us at McKinsey.com to get the full report, A guide to impact investing in Black economic mobility. Shelley, thanks so much for joining us today.
Shelley Stewart: Roberta, thank you very much for having me.
Lucia Rahilly: Next up, many AI projects get stuck in a lab. Asaf Somekh is the CEO and cofounder of Iguazio, a data science company that automates machine learning pipelines for real-world outcomes. He spoke with Roberta about how to dislodge those innovative projects and move them into the real world.
Roberta Fusaro: Asaf, welcome to the podcast. It’s great to meet you.
Asaf Somekh: Thank you, Roberta. Happy to be here.
Roberta Fusaro: So we’re just on the other side of Mobile World Congress—this huge annual event focused on connectivity technologies and providers and platforms that you attended. What are your biggest takeaways from the event?
Asaf Somekh: I’ve been to many MWCs before, and this time there wasn’t much hype or people just talking about 5G. People were there to see how they could improve their business’ top line or improve their cost, things like that. It was back to normal business. People need to focus on profitability.
Locked in a lab
Roberta Fusaro: Speaking of hype, a lot of the hype nowadays is around generative AI. How does your work with applied AI and MLOps [machine learning operations] intersect with work being done on generative AI?
Asaf Somekh: Generative AI has caught the attention of many, like ChatGPT and related technologies. But once again, people are looking for practical implications for their business and to understand how to apply the technology. We had great conversations around improving your call centers when it comes to the large telcos, or on the campaign side, you can use generative AI to develop the campaigns or the texts that are specific and personalized for their customers. I mean, huge potential. But people want to understand better how to go about it, and what would be the infrastructure in place to actually make it practical. Because at the end of the day, there is a huge gap between the work that has been done with AI in the lab and the current value that it brings to the enterprise as a whole. This is a big gap that exists today in the market.
There is a huge gap between the work that has been done with AI in the lab and the current value that it brings to the enterprise as a whole.Asaf Somekh
Roberta Fusaro: I know there’s McKinsey research showing that organizations worldwide have invested almost $500 billion in applied AI, but only 10 percent of AI projects actually emerge from the lab.
Asaf Somekh: Yes. So just as an anecdote, one of the clients that I met used the phrase that I love to represent this inefficiency in the market. She said she had “pilot sickness, because I’ve done so many pilots with AI, wonderful pilots, but what about results to the business?” So that’s exactly the problem we’re trying to fix.
Machine learning operations may hold the key
Roberta Fusaro: What role does MLOps play specifically in addressing this notion of pilot fatigue?
Asaf Somekh: MLOps takes a production-first approach. In the lab, you’re thinking forward on how that will play out in production. MLOps is about the process—of starting in the lab and making sure that whatever is developed in the lab can scale when it comes to the daily use of the application in production. It should also have the ability to monitor predictions, by integrating changing data patterns or the demography of a particular use case. Opposed to general software, models are things that need to be constantly retrained, and MLOps has to identify the need for retraining these models and apply that while you have the entire pipeline of applications running at scaling production.
Think about it as a machine learning factory as opposed to an artisan that builds a nice chair in your garage. It’s nice, but if you need to build a thousand chairs, you need to have a factory. And that’s exactly what MLOps is all about. It’s to create that factory within the enterprise.
What’s getting in the way
Roberta Fusaro: What are some of the obstacles organizations face when they’re trying to roll out and scale up AI experiments?
Asaf Somekh: We see many silos today. You have a silo of data scientists that are just doing research and then the next silo of data engineers that are taking the result of that and trying to re-create it. Sometimes we see data scientists working in Python and then data engineers working in Java and things literally get lost in translation because they’re speaking different languages.
AI is still all about software, and it has to be integrated with the general practices of software within enterprises. That’s one of the obstacles we see today, where data science is completely separated from the general life cycle of software, and sometimes the work is thrown over the fence to people that need to translate it to the general software machine learning—and that doesn’t work well.
Roberta Fusaro: Are there examples of companies starting to get this right or are there particular industries where we’re seeing applied AI succeed more or less than in other sectors?
Asaf Somekh: So obviously, the tech giants have been leading that. I call enterprise AI today a luxury of the tech giants because they understood this gap. They invested heavily with hundreds of people in R&D teams that built their entire infrastructure to address these problems. But most enterprises cannot afford to do that. It’s all about productizing machine learning in the organization. This is a category that has been evolving over the past few years. You have products and platforms and services around this technology, and Iguazio is one of them.
The ethics of AI
Roberta Fusaro: What are the ethical considerations that companies need to think about when using applied AI? Are there resources—a sort of AI bill of rights or something—out there for organizations to incorporate into all of their experiments with AI?
Asaf Somekh: It’s a wonderful question because we don’t, in fact, need to come up with it. As usual, when it comes to regulation, it starts in Europe. Europe is already in the process of formalizing AI regulation. My recommendation to many enterprises is be prepared for that. You need to have the infrastructure in place. Often people start the AI journey by only thinking about developing models in the lab and just plugging them into their infrastructure somewhere. It’s way beyond that and why we’re here.
Roberta Fusaro: What are the talent requirements for this particular technology?
Asaf Somekh: What we see today is that machine learning engineers and data scientists are spending too much time on infrastructure. Once you bring in the MLOps platform, those people can actually focus on the business problem, on developing the actual application instead of dealing with the pipelining and the infrastructure around it. That’s part of the problem. Adding to that, companies hire really expensive or hard-to-find talent to solve the business problems, and they’re spending only 15 percent of their time on the actual machine learning and data science work. The rest of their time is spent on the infrastructure. This is another key problem in this space, and it creates frustration.
Roberta Fusaro: Thinking more broadly, what excites you most about the future of AI?
Asaf Somekh: The whole holistic approach where every transformation that is happening now is including AI in its core, and the ability to deliver complete solutions by vertical markets. That’s when we can actually help companies leverage AI, and bring the potential of AI to reality.