Esteve Almirall, Professor, Department of Operations, Innovation and Data Sciences at Esade.
Many of you will surely remember the "Mobile First" era, a phrase that defined a decade. Mobile First aimed to describe a paradigm shift. Until then, applications, processes, and user interactions were designed with the web in mind. Mobile First suggested a pivot—designing for mobile, for users who were no longer seated in front of a computer but were perpetually connected via their phones, a practice that quickly became the new normal.
So, are we at such a pivotal moment with A.I. now?
For a long time, A.I. was synonymous with optimization or automation—in other words, replacement. Yet, the process design remained structurally the same. For instance, human critics were replaced by software for making recommendations, swapping out components as if they were Lego pieces.
It's true that industry leaders have long since moved many elements of coordination and even the entire process into code. This isn't a new topic; Uber coordinates "drivers" and "riders" automatically with A.I., as do most platforms. Similarly, companies like Amazon have their process management and coordination, including logistics, in code. However, most companies design their processes and workflows placing A.I. in the realm of optimization.
This approach has its obvious advantages, such as less complexity. There's no need for any process reengineering or to tackle organizational changes; everything is, in essence, simpler.
But with the advancements in A.I., isn't it time to emulate tech sector leaders and reap similar benefits?
The "how-to" doesn't differ much from most innovation and reengineering processes: copy, adapt, and innovate.
A fundamental shift in attitude is required; you're not that special, and the world's brightest minds aren't all on your team. Thus, the first step is to learn what the best are doing. If there are examples similar to what you do, then great; if not, it's about transferring what the best do in other organizations to your business, to your team.
In essence, it's about knowing with as much certainty as possible where the technological frontier of your activity lies and to what extent your capabilities allow you to be at that frontier.
The second point is to adapt and remix. Innovation arises from transferring new discoveries, technological or otherwise, into a specific environment and making them viable with a satisfying business model—fortunate enough for users to adopt them. It also arises from recombining known and tested proposals with existing ones. From this recombination sprouts something new and distinct, which, by evolving and adjusting to its environment, may become the innovation of the moment that everyone will rush to copy.
Incrementally innovating the resulting proposition is crucial, because an innovation is only such if it is adopted by the market and it will only be adopted if it fits.
There are two more variables in this process. The first is how much risk you're willing to take. It's not about creating the most groundbreaking product/service; it's about constructing the one that will be most adopted and improve our market position.
For this, you must consider how much you stand to gain. This is always a difficult question because it's situated in the future, it's somewhat about predicting the future, which is never easy, especially when done from the past!
A tip: ask yourself how you compete and how you could compete. You can do this at the level of your organization, your team, or even yourself because, in the end, we all collaborate and compete to some extent.
The last variable you must control is the possible evolution of both the technology or the model you're going to implement, as well as your competitors. Remember, we are neither alone nor is the world standing still. It's not about being first, but about finding the right time, the appropriate moment.
Thinking about these types of restructuring, as in A.I. First, involves learning and therefore developing pilots. Let's look at a couple of examples of how first-rate companies outside the tech sector are doing it.
One of the most promising fields is undoubtedly automating "Customer Service." We all foresee that generative A.I. tools like ChatGPT will manage customer complaints better than current operators, but it's a high-risk field, very public, very exposed, and there are few experiences. Some organizations like McDonald's or Delta are experimenting with internal services resolving their collaborators' doubts. It's about learning while minimizing risk and having space to adapt.
In the same vein, other organizations are following a different approach. For example, Gucci is piloting the same technology but instead of replacing, they use it to augment their collaborators' capabilities by doing cross-selling and resolving their problems better, achieving a 33% increase, which is no small feat. It wasn't something they sought, it was something they discovered by experimenting.
Of course, you can always wait, then all these doubts will be resolved, by others... the risk is high, will you still be in the market?