Rise of AI: Is Your Company Prepared for Generative AI?

There’s no need for business leaders to panic about the arrival of technology like generative AI. Wharton’s Rahul Kapoor explains why now is the time to develop new frameworks to manage the changes ahead.

Dan Loney: There’s such a large conversation going on right now around open AI, ChatGPT. When you think about management and companies, a lot is still to be determined, but it seems like there is great potential.

Rahul Kapoor: As somebody who has done research on deceptive technologies and business models for almost 20 years, ChatGPT, generative AI more broadly, is truly emerging as what I would call a general-purpose technology, like semiconductors in the ‘50s and the ‘60s, like the internet in the 1990s. Clearly, they are emerging across multiple use cases, multiple applications. There is what I call an “era of ferment” that all of us are in, where there are multiple use cases, multiple business cases, multiple application domains, with enormous potential for growth and disruption across the economy.

How Much of a Disruptor Is Generative AI?

Loney: Would it be safe to say this a big pivot moment for a lot of these companies?

Kapoor: If you look at semiconductors and internet as two of those disruptor forces that we have lived through, what is similar between those and generative AI is that they are solving two types of problems. One is a problem around productivity/efficiency. Can we do tasks more efficiently, whether it’s time or cost? As you know, internet and semiconductors really pushed the frontier around those issues.

The second is, can we find new ways to create value? Whether it’s in the enterprise side of things or the consumer side of things. Generative AI seems to check both of those boxes. If you think of many markets and professions, generative AI is going to make things more efficient. At the same time, it’s going to present many cases of additional value being created through those. I do think this part would suggest that this is not a three-year or a five-year phenomenon. This is a generation-wide phenomenon that we will all be living through for the next 20, 30 years.

How Are Companies Adapting to the Rise of Generative AI?

Loney: What are the challenges when you have those massive changes? How does management and how do corporations react to that?

Kapoor: It’s a commentary that I’m seeing more and more coming through in the popular and the business press. There’s clearly discussion about different types of use cases, different types of industries, different types of professions that are getting impacted or going to be impacted very significantly. Things like customer services, content creation, software development, internet search, data analytics. There’s enough evidence that they are going to be impacting them in a pretty short order.

Then you have what I call recombinations. These are not markets where there’s an existing way of doing things and this new technology can do it better. Recombinations are where new use cases emerge through those new technologies as building blocks. Think about ChatGPT, a combination of virtual, augmented reality, as a completely new use case that you and I have not really recognized before.

What leaders and managers can do is recognize that these paradigmatic shifts don’t take place in two years, three years, five years. The internet emerged in the 1990s. Amazon, as the poster child of the success of the internet, was successful 20 years later. Semiconductors emerged in the ‘60s. Intel, as a poster child for a semiconductor leader, was only visible in the 1990s. The good news is that this is not going to be a disruptive change that’s going to happen overnight. That means there is no need to panic, if you are a leader, if you are a manager.

But there is a clear and critical need to understand what generative AI or ChatGPT can do in terms of your business. This is where most of my research and teaching at the Wharton School engages in terms of a framework that leaders and managers can use. I’ll just give you a very short aspect of that framework. If you are a leader or even an entrepreneur in an emerging business, you have to think about the assets in the business market that your firm is engaging with. And then you have to think about these shifts taking place in the technology landscape. ChatGPT, for example. What aspects of your assets and business models could complement what ChatGPT could do?

If you are in the health care services, maybe there are certain things that would continue to be done the way they are done. But then ChatGPT or generative AI becomes a value-added service for the patients, for your clients. If you are in the content creation business, there are certain aspects of content creation that could be fundamentally replaced through generative AI models. I think that’s a starting point, how your existing assets and business model are going to be shaped. And not framing it as a negative threat, but think about where it can complement and where it might replace.

A very parallel thought process that I engage with students and executives is taking on the customer side. You have existing customers, then you’re looking for growth as new markets and new customers and existing customers are part of an existing ecosystem. New customers may require you to build a new ecosystem. Once you start thinking about the ecosystems, you think about how a new technology like generative AI affects your existing ecosystem or your partners? Are your customers better off, or do you need to invest in the ecosystem to create a higher value for your customers? Do you need to create new ecosystems, like electric cars? What Tesla has done and the success story there is that it’s not just about the electric car. It’s about building the whole ecosystem.

The organizations and the leaders who are going to do well are the ones who are going to have a very thoughtful approach to how this shift is affecting their business, both in terms of the positives, but also in terms of the challenges. Thinking about it not in terms of just existing customers, but new market opportunities. And not just focusing on the existing value chain, but the broader ecosystem of how these companies create value.

How Will Generative AI Impact Consumer Relationships to Business?

Loney: There’s been so much conversation in the last few years about the relationship between the company and the customer, about how the industry can do a better job of making that connection and keeping that consumer longer term. Is that the next step in this process?

Kapoor: We’ve talked about customer lifetime value. As we have embraced more personalization, much of this is coming through data and data analytics. We are explicitly thinking about how we can enhance and maximize the value that we can create for every customer. It doesn’t require a one-size-fits-all approach. We can generate things that are more personalized. We can create and capture value for each customer or entity over time.

I do think a generative AI would enhance that. It would broaden the scope of what personalization might mean. For you, personalization might be from a content creation perspective. For me, it would also include the research perspective as well. I can imagine scenarios where companies are able to create a suite of personalizations and bundles through generative AI at the level of a customer, at a level of enterprise that was not possible before, primarily based on data analytics. I think once we include these new algorithms and these new learning processes that are algorithmic, the scope of personalization becomes much broader. And that has exponential effects in terms of success of businesses.

What Will Scalable Implementations of Generative AI Look Like?

Loney: You can even take it larger scale and think about how it’s going to potentially impact whole industries, correct?

Kapoor: Absolutely. In fact, I was on a panel this morning with fellow academics from other institutions, and that was our question: How can we think about the impact of a general-purpose technology like generative AI? The conversation that I strongly believe is true is, it’s going to have at least four different points of impact. The most obvious one, as a strategy professor, is for businesses and for industries. But it will also start having an effect on professions and societies at a much broader level. Think about what social media has done to our societies. Think about what the internet has done to our society.

I think each of these levels, starting from organizations to industries to broader professions and even the social fabric, are likely to be impacted through this technology emergence. Some of this is going to be positive and value-creating, and some of it we have to be careful, especially with such a strong human interface and a modality that these technologies are human-like. There has to be care that needs to be exercised in terms of how professions and societies are going to be impacted by these emergences.

Loney: Part of that is how companies think about their structure and the types of jobs they will have.

Kapoor: Absolutely. There is a lot of research. It’s very early stage, so we don’t have a definitive statement on that issue. But the model that we all prescribe to as researchers of technology is that every technology follows an S-shaped curve. The early stages of a technology, it’s imperfect. It’s not cost-effective. We have to invest in resources, and it takes time for it to improve.

But once the technology gets well understood and experimentation allows us to fine-tune how the technology actually works, then we have a very fast, robust trajectory of improvement. Kind of what we saw in semiconductors in the ‘70s or the ‘80s, what we saw in the internet, post-2000, what we’ve seen with electric cars in the last three to five years.

Many different professions are going to be affected. I think it’s a foregone conclusion. The rate and scale of what that would look like would be a function of those S-curve trajectories across the different professions. Let’s think about the professions that are tied up in customer services versus professions that are tied up in creative industries. You’re likely to see different models of S and different rates of growth, depending on how the technology is going to improve and achieve the performance requirement for those professions to be substituted by these emerging modalities.

Loney: My thought is that the sheer volume of companies and industries that could be impacted by this feels extremely large. And it will be on the companies themselves, or maybe the industry, to understand how Open AI, ChatGPT will be incorporated within the structure of their firm.

Kapoor: Absolutely. And just the way you describe the problem, it’s a very complex problem, right? When I talk about how companies and leaders can manage disruption of a scale like what we’re talking about today with generative AI, there are three impediments that companies have to deal with in the short term. First, there is significant uncertainty in terms of what the dominant use cases are going to be. What solutions are going to satisfy those use cases? What business model would allow for companies embracing generative AI to actually do it in an economically viable way? There are many sources of uncertainty that every era of disruption seems to present. You are in an environment where there are more unknowns than knowns.

Secondly, you still have to run your existing business. This is going to be a slow process of evolution — three, five, many cases longer, years. How do you want to still focus on the old business or existing business while exploring these new possibilities. And the third is the adjustment cost that you are making in terms of building new skills, building new capabilities, dealing with new competitors.

Once you introduce inertia, which is how you change, once you introduce uncertainty, and once you introduce the balance between the existing and the new, it is a very, very hard nut to crack. I’ll tell you what I’ve seen in my research. The best time to explore something that’s uncertain and disruptive is when you’re not desperate for it to be adopted. The best time to invest in a new disruptive solution or opportunity is when you are running a healthy business. And that’s where the pressure is going to be the lowest for you to take more risk and explore more broadly.

The second aspect of the management or leadership that generates successful outcomes is to think about what sort of organizational structure would allow for companies to take advantage of these disruptive opportunities. It’s often talked about that they need to be managed as separate businesses or separate units. You see what Google is doing with Alphabet. There’s a Google business, and then there is a Waymo business, which is a self-driving technology. I think that’s one way to do it.

Partnerships, alliances with many companies to share risk, to share cost, to learn from each other, is a model that I think is a very effective, especially when it’s still uncertain. Third, I think we need patience while we need to be disciplined in terms of how we experiment and innovate as well.

I think it’s a combination of, how do we balance existing with new, how do we engage in an organizational configuration that allows us to maintain that balance, and then having a disciplined approach to how we are experimenting around these new technologies and business models, and giving them enough of a runway to take off as opposed to switching back and forth between something new versus something else every year, every two years. That generates more confusion, more inertia, more cost, and tends to be counterproductive.

Is Generative AI Just Another Game Changer or Something Much Bigger?

Loney: We have had disruption in our corporate structure for such a long period of time with various impacts. But this feels like disruption that is taking us down a new path. It’s almost disrupting what we think the idea of disruption is.

Kapoor: I think the jury is out, in my view. If I might use this analogy: Is this is an old wine in a new bottle? Whenever there is a widespread paradigmatic change like the internet, like semiconductors, those conversations emerge. I see it with generative AI or ChatGPT. I’m seeing there are those two camps within academia as well. Many people believe that, yes, the technology is different. The models of value creation are different compared to what we saw with the internet or semiconductors. But the nature of shifts in terms of assets and business models and capabilities have been well understood and studied. So, it is a wine that we understand, and we can understand what the ramifications are. Others believe that this is fundamentally different. It’s disrupting the way we have thought about disruption.

I think how to parse these two perspectives depends on thinking through what the impact that generative AI is having, in a business context. It’s clear that this is not just a technology problem. It is a broader issue than just thinking about the technology. It’s how it impacts organizations, in terms of the jobs, but also in terms of how organizations create value. It impacts the business models. It can completely replace a different business model and find a completely new business model at the same time. It also impacts the broader ecosystems that many companies are part of, in terms of requiring a major reconfiguration of the ecosystem or making many old ecosystems completely redundant.

On top of it, generative AI technology is allowing us to do two things simultaneously that has been very hard to do historically, which is that we can use generative AI to find use cases that create more value for the specific customer or market. At the same time, we can do so at a much lower cost. I think because generative AI can simultaneously help businesses and customers to create more value at lower cost, the effects are going to be more exponential than linear. That last part suggests the theory that this is new form of disruption that we have not seen. It starts looking more promising. Not so much that we have not seen a flavor of that before; I think internet had flavors of that in very specific markets like social media and others. But I think it’s not happened at this scale, across markets, across industries, across professions, across societies, at the pace that we think generative AI is going to propagate.

I think the scale of the effect and the linkages between technology, organization, business models, ecosystems, and professions are going to create a much larger multiplex effect. Some of it we have understood and seen before. But some of it remains to be studied and remains to be understood.