Why AI shouldn’t change your business strategy

If AI changes your business strategy, then something was wrong with your business strategy.

I was asked by GSMA to deliver a keynote this week as part of its Telco AI Summit held at Mobile World Congress Las Vegas. The summit program explored how AI and GenAI are being used by operators, how usage will evolve over time, and the implications for performance, customer relations and competition.

It was encouraging to hear grounded discussion amongst the participants.

Often when discussing promising new technologies, we focus on the wonder of all that is possible. That’s great for creating buzz, but not so much for delivering results.

In reality, when we consider AI’s capacity to power transformation, we should remember that it shouldn’t fundamentally change anything about our business strategy. If it does, it means we weren’t doing the right things in the first place.

AI should merely help us do things better, faster and at scale.

Rakuten runs the biggest loyalty program in Japan. The business strategy is to leverage this loyalty across the company’s 70+ businesses to drive synergistic growth. As a tech company doing telco business, we have reinvented customer experience and efficiency, disrupting markets along the way. This remains our mission. So when we evaluate AI or another technology, it is solely through the lens of how it can advance the business strategies we already have in place.

What’s most interesting about AI

The interesting part about AI is not its many benefits, which have been covered ad nauseum. It’s not the unprecedented risk that it accompanies it. Usually, when these aspects of AI are being discussed it’s because someone is trying to sell us a solution for the former or for the latter.

What is most important about AI in my view is what this technology means to us and how it works. Because if we understand this, we can decide how we want to use AI.

All technology is neutral. It’s what you choose to do with it that matters most when designing tech-powered strategies.

Yet, our path in telecom tends to diverge from reality when new technologies emerge. Sometimes we stray temporarily. Sometimes we get to the very end of the path to find we’ve gotten nowhere.

This happens when we attach an almost mythical belief to technology we don’t understand. As humans, we’ve done this for centuries.

Destroying the magic of AI

There was a time when parts of civilization prayed every morning to ensure the sun would rise because people didn’t understand how the sun worked.

If we don’t understand how AI works, we may end up praying to it in a similar way, which will lead to poor business decisions. We need to understand what it can do, what it can’t do and how it does anything. Then we understand the complexity and value of any solution being presented to us.

You can outsource solutions, but not understanding

AI is incredibly mechanical—it predicts the next word in a sentence based on probability. But it feels human and almost genius-level because of the vast amount of data it can process and output.

We are witnessing the industrialization of creativity.

As humans, we contend that it is our creativity that separates us from other forms of life. But the vast majority of what we do is not based on original thought. Original thinkers like Einstein or Aristotle are extremely rare. We know this because we name original thinkers on the fingers of one hand and we all know who they are.

This means most of what we call creativity can absolutely be duplicated by the approach of GenAI since it is reworking what already exists into new forms. This feels creative to us because it is what we are used to doing.

When we refer to creativity here, we refer to any profession that uses words and images to create additional versions of words and images. Under this definition, the legal profession, for example, is creative and GenAI is proving to be highly effective in creating legal documents at a fraction of the human price and lead time.

Techo versus telco

The AI rubber is already hitting the network road.

When we're looking at the application of this technology inside our business, we have to decide if we're looking at it from the perspective of the technology and how we deliver it or the end user and the experience we want them to have.

This is the fundamental difference between a techo and a telco.

A techo doesn’t start with a new tech capability, but rather, a perspective on how to change the existing world to be better for the customer and themselves. That's why for an internet company, AI is nothing new. It is just oxygen to fuel the existing fire of a data-driven organization that allows them to do things that couldn't be done before.

In telecom, we inherently understand technology and delivery. Where we’re struggling is finding growth that can accelerate our journey away from where we are today and to hyper-competitive businesses that thrive in what we know is a changed environment. Once we have answers here, technologies like 5G and AI will make sense.

Buy or do?

The question we have to ask ourselves then is do we buy or do we do when it comes to AI?

The thought process is no different from what we do with any other technology, such as cloud.

By 2033, telecom will spend an estimated $24B on AI software and services, with an expected CAGR of about 30%.

If we are using AI to increase efficiency, existing costs need to be reduced at a greater rate than the new spend being made to reduce them. It is important that stakeholders evaluating the myriad solutions available to them understand what they believe they need to buy. If they know what they need to deliver on an outcome (i.e., the techco approach), they will understand if they’re underpaying or overpaying and be able to control their AI future.

There are early use cases where AI can be highly supportive of a desired outcome:

• Toward a sustainable future. In telecom we’re sustaining an incredibly large cost structure that we know is not optimized. Our networks are on 100% of the time, regardless of whether they’re being used. This was once necessary because it was difficult to switch parts of the network on and off without disrupting service. But with AI, we now have the data granularity and precision to manage network resources at more specific levels—such as cores, sectors or antennas. Even a small reduction in energy use, like 1%, can lead to significant savings. Yes, this speaks to the power of AI, but as said earlier, AI is only as good as the data it relies on. To use AI effectively, we need to ensure our data is trustworthy and available in real-time. Imagine your network as a system of light bulbs. Which lights should be turned off when not in use and how quickly can they be turned back on? Operators that are able to manage this at a granular level will see major improvements in sustainability and energy efficiency.
• Good versus evil. Because AI is a neutral technology, it can be used for both good and evil. Already, its potential for malicious use is being exploited. Personalized phishing attacks are becoming more sophisticated as attackers no longer need to cast wide nets. Rather, AI helps them create highly targeted, individualized attacks at scale and at zero marginal cost. AI presents opportunities for telecom companies to shift focus, offering safety and security as part of the value we deliver to customers. This is especially important in a world with an aging population increasingly vulnerable to online threats. As telecom providers, we have a unique opportunity to provide solutions that protect these individuals and deliver something far more valuable than just a network connection.

Focusing on what matters

AI is nothing more than another software algorithm—except that the complexity has shifted from the code to the data. Now, we must manage and curate that data to efficiently deliver results.

AI allows us to operate at a different scale, with new cost structures and faster timeframes, but it shouldn’t fundamentally change the business’s core mission. The journey is not about quick wins but grinding it out over the next decade and becoming proficient. The organizations that succeed will be the ones that understand AI deeply and apply it thoughtfully, making incremental improvements to processes along the way.

Mention me Geoff Hollingworth in the comments to share your view or start a conversation.