AI: The case against use cases

To embrace AI disruptive innovation, stretch your imagination beyond your existing business model.

There is an extraordinary interview between BBC journalist Jeremy Paxman and David Bowie—one of the most influential musicians of the 20^th century—from way back in 1999. Paxman’s assertion is that the nascent internet is “simply a different delivery system,” merely an alternative way of doing something that already exists. Meanwhile, Bowie states that we are actually “on the cusp of something exhilarating and terrifying.”

At the time this interview was recorded, I was working for a government department’s intranet site. My main role was to convert 80,000 Word documents to HTML. Suffice to say, my thinking was more aligned with Paxman’s than Bowie’s.

Thankfully, I have moved on since then, and—like many—I recognize that the growing adoption of artificial intelligence (AI) once again puts us “on the cusp of something exhilarating and terrifying.” But I also believe that many established organizations are failing to recognize both AI’s potential and the extent to which it may represent a threat to their very existence.

The difference-maker lies in these established organizations’ perspective on innovation. There are many ways to categorize the different types of innovation, but from an organizational perspective, we can classify these as either working within an organization’s existing business model or operating beyond it. For those organizations willing to make the bold move of espousing the latter type of innovation, the full benefits of AI await.

AI innovation within the existing business model

First, let’s look at the two main types of innovation that can operate within an established organization’s existing business model:

1. Efficiency innovation: Doing more with less. This entails finding ways to reduce costs, often without customer awareness of the initiative. For example, many organizations are now using AI as “a different delivery system” (to quote Paxman) for their communications—a cheaper, more efficient way to deliver an existing service that customers are (hopefully) unaware of. McDonald’s adopted this approach when it introduced AI as a means of automating its ordering system.
   
   
2. Sustaining innovation: Extending your organization’s current product or service offering. The customer is usually made very aware of these increased service offerings (perhaps sold at a premium), which are introduced to protect or expand your market share.
   
   An example is organizations incorporating AI into a service offering in order to expand or extend it, such as when Apple announced it would allow its iPhone and Mac operating systems to access ChatGPT.

Both of these opportunities are relatively easy to identify and manage. Teams and individuals are often keen to pinpoint “use cases,” and funding can be readily acquired for business cases that are aligned with existing operational success criteria. Even when the forecasts are rather optimistic, the criteria are familiar enough for stakeholders to imagine a scenario in which they might be achieved. Confidence is high that teams will seamlessly be able to adopt this technology, as they work within the comfort of their organization’s tried and tested business model metrics.

At best, both these approaches improve or protect the organization’s bottom line. But at worst, they can reflect a desire to be “seen to be innovative.” Inappropriate use cases are identified, and processes are automated ineffectually (see McDonald’s bacon-topped ice cream) or products are extended to the extent that the customer is over-served (see Apple’s drop in share price). 

AI innovation beyond the existing business model

Now let us consider two types of innovation that typically exist beyond an established organization’s business model.

1. Disruptive innovation: Where a rival company either enters the market at the low end with a “good enough” product (before moving incrementally up-market to acquire the incumbent’s customers) or creates a new market by offering a simpler, cheaper or more convenient product to a previously under-served segment of customers.
   
   
2. Radical innovation: Where new technology opens up new markets, potentially redefining—or even creating—an industry. 

These opportunities are far more difficult to identify by established organizations. This is because established organizations are configured—through their business model and success metrics—to train their focus on their existing customers, not to seek out underserved sections of their market. As described by Clayton Christensen in his influential book The Innovator’s Dilemma, they are not motivated to disrupt their business by defining entirely new markets.

The exploration of potential new industries enabled by radical innovation is also inadvertently discouraged, as teams are required to work within an established business model. 

These organizations’ competitors and future competitors, however, are not shackled by such constraints. Last year, our colleague Duncan Roberts discussed the emergence of “AI-native organizations,” for which AI will be as fundamental as electricity or the internet is to established organizations. This will enable them to create simpler or cheaper products and services that will gradually disrupt incumbent organizations. Or they might use AI to provide a service to an underserved market segment in a way that makes it affordable and accessible. Or they could identify whole new industries beyond the bounds of our current understanding of society’s wants and needs.

Preemptive business disruption

So, how can established organizations respond to this imminent market disruption?

After reading The Innovator’s Dilemma, Jeff Bezos famously chose to disrupt his own business model, before it could be disrupted by a competitor. When he moved Steve Kessel from managing Amazon’s physical book category to the company’s digital efforts, he informed Kessel, “Your job is to kill your own business. I want you to proceed as if your goal is to put everyone selling physical books out of a job.” Kessel’s team went on to create the first Kindle.

Few have the authority (or corporate culture) to mandate such a directive. But it is essential that when considering disruptive and radical innovations, we recognize that gauging these initiatives’ merit through the prism of our core business’s established success metrics—or, indeed, fundamental business model—can prevent us from identifying their potential.

Looking beyond the business model for disruptive AI innovation

Businesses should adopt processes and foster a culture that evaluate their AI initiatives beyond the confines of their established business model—but this is easier said than done. Within the core business, we are all incentivized, rewarded and promoted on the basis of criteria that are aligned to this business model—which has previously served the organization well.

Evaluating a new technology as a way to simply do what we currently do more efficiently makes perfect sense. However, as the authors of the book The Other Side of Innovation argue, “Innovation and ongoing operations are always and inevitably in conflict.” 

We need to lean into this healthy conflict. One way to achieve this is to establish small, dedicated teams to explore and develop new concepts and opportunities. These teams can target new revenue streams, customer cohorts, distribution channels and indeed business models, that can help the organization mitigate and anticipate the disruption AI will instigate. Crucially these teams must:

1. Identify problems to solve, not user cases to exploit. The team should spend time with customers and users to empathize with them and glean insights.
   
   
2. Be unshackled from core business model metrics. Instead, they should be empowered to define their own success metrics, with an emphasis on “validating learning” as they seek to achieve market fit — if not market creation.
   
   
3. Look beyond their organization’s core competencies. This introduces the opportunity to partner with AI startups.
   
   
4. Be conscious of the social responsibility of their solutions. Typically, innovation teams focus on the desirability, feasibility and viability of their products and services. But with AI, it is ever more important to consider not just whether we can create a product or service, but whether we should.

 It’s not easy to get buy-in for disruptive innovation, nor to establish a distinct culture of experimentation and validated learning. But if you solely consider AI from within the confines of your established business model, you risk adopting Paxman’s perspective that it is just another way of doing something that you already do, rather than espousing the enlightened opinion voiced by Bowie that the opportunities are barely imaginable—at least by your own organization.