The AI bubble will burst fast, then the real work begins (Reader Forum)

Perspectives from MWC24 and where we go from here

You wouldn’t know it from the 12-foot high letters adorning booths throughout the halls of Mobile World Congress, but the AI bubble is about to pop.

Soon, it won’t be enough to simply be “doing” AI or for some new tech feature to be AI-powered. Every stakeholder will need to demonstrate progress toward a desired outcome that can positively move the needle on telecom’s long standing challenges: viability, profitability and market competitiveness. Yes, the same problems that existed before we started collectively blowing hot air into this AI bubble. 

With that in mind, I wanted to share thoughts on developments coming out of this year’s MWC and reinforce a couple critical imperatives so that when this bubble pops we’re all ready to dig in and do the hard work that’s been sitting, waiting for us.

AI is not something you (should) buy

Vendors loudly proclaimed their readiness to sell “AI.” But there is no AI unicorn that will solve telecom’s problems. 

When we go shopping, we don’t set out to buy a “software-powered” car or headphones or printer. 

Remember when IBM made a killing selling storage products to big companies that were convinced they needed storage before they even had a use case in mind? Sales eventually collapsed when these organizations realized storage wasn’t the answer to any of their problems. 

“AI” doesn’t describe any form, function or value. It’s one of the fastest destructions of a meaningful word I’ve seen. 

AI is cheap. But if you try to buy success, you will pay a premium and have to know how you will get that money back. Inevitably, you may end up funding someone else’s success. We have to focus on outsourcing solutions, not problems. It’s not about the size of the LLM. We don’t need to replicate the total intelligence of humanity to help telecom. Understand focused outcomes and apply AI to them. It will quickly get cheaper and cheaper to ride a new wave of improvements.

The fallacy of AI alliances

Finding like-minded stakeholders that share your vision and want to march arm-in-arm toward the future can feel good. It wouldn’t be MWC without some fresh alliances. But despite the fanfare and optimism, there’s scant evidence these initiatives can actually make strides toward solving real-world problems.

I’m confused when I see so many players joining to inject AI into the RAN. It wasn’t long ago the industry was abuzz around the Radio Intelligent Controller (RIC) and the prospect of standardizing data output from radio networks. We only have one radio network. Why are we forming more alliances that, while well-intentioned, risk diluting focus and diverting valuable resources from where they’re needed most?

Over the past decades, I’ve observed that numerous alliances, despite initial hype, have failed to materially impact business outcomes or just quietly faded into the background. The question then becomes: what if the energy and resources devoted to forming and maintaining these alliances were instead directed toward understanding and catering to customer needs? 

It’s a proposition that suggests a potential recalibration of priorities—shifting the industry’s focus back to its foundational purpose of serving customers and addressing the tangible, day-to-day challenges and opportunities that directly impact viability, profitability and market competitiveness.

It starts with automation

The transformation we’ve all said we want starts with simple automation and rules-based closed-loop systems. It is marked by a significant transition in organizational governance, moving the locus of trust from human expertise—relied upon for experience and knowledge—to software and data. This means a real difference in how we perceive and utilize technology within our operations.

This transition is not merely a technical adjustment but a paradigm shift that affects every level of an organization. The experience of Rakuten in deploying ‘industrialized digitalization’ within its mobile operations highlights an essential aspect of this journey: the social engineering required to maintain confidence in these automated systems. Ensuring that experts have control over when to trust the system combined with a clear chain of custody for actions taken by automated processes, lays the groundwork for a secure and responsible AI and automation infrastructure. This approach allows for accountability and rectification processes, such as post-mortems, when things deviate from the expected outcomes.

What Rakuten’s experience teaches us is the importance of securing the automation and AI supply chain—a responsibility we bestow upon our software. It’s about building a framework within which automation can be trusted at the same level as human decisions, if not more. This shift necessitates a careful balance of speed and control, ensuring that as organizations move quickly to embrace automation, they do so with the full buy-in and momentum of their entire team.

As we navigate this path, the lessons learned in the initial stages of automation will inform how we scale these technologies responsibly, ensuring that they serve to enhance rather than replace the human element within our industry.

Runners, take your marks

The strategic imperatives of telecom haven’t changed. We need to increase our effectiveness and efficiency at running networks in a more agile manner. If you’re looking for an indication of our current misalignment, just count the number of technology references at MWC versus references to customers and what we are delivering in terms of what people want. It’s hard to recall any conversation where the customer was the main actor.

However, amidst this misalignment lies a monumental opportunity, underscored by the latest developments in AI. We are seeing a paradigm shift toward software not just as a tool for automation but as an integral part of solutions, capable of making decisions and processing. This evolution of software’s role will change how software interacts with organizational workflows and processes.

If we change our focus to truly leverage these advancements, we change our collective fortunes in the process. The impending pop of the AI bubble will indeed be loud. Let’s take our marks now, consider this moment as the starter pistol’s and sprint towards a future where success is defined not by how well we tout technology but by how effectively we harness breakthroughs to meet and exceed customer needs.