Interview published in Forbes Spain – November 2024 – Download original PDF in Spanish

—–

We know it is a technology that will change our daily lives, but are we prepared for the democratisation of artificial intelligence?

Value creation, automation, enhanced competitiveness, organizational shifts, and regulatory challenges—as with anything, Artificial Intelligence has its benefits and drawbacks. Especially now, as it has moved from the realm of science fiction to firmly take root in our lives with the rise of Generative AI. This is why we spoke with Juanma Aramburu, CEO of the tech consulting firm Keepler Data Tech and an expert in maximizing the potential of this transformative technology.

Question: What stage are we experiencing with AI right now? What would you compare it to?

Juanma Aramburu: I’d compare it to the Industrial Revolution because of its transformative impact on work and certain professions. But more than that, it feels like the arrival of mobile phones. Initially, only a few people had one because they were expensive, but they quickly became essential, to the point where it’s now unimaginable not to have one. AI has been around for a while, but it’s now accessible to everyone, especially with the advent of Generative AI.

Q: Does it really make that much of a difference compared to existing AI?

JMA: In traditional AI, use cases are usually focused on a specific problem or challenge, and the return on investment (ROI) is relatively easy to measure. For Generative AI, however, measuring ROI is more complex because its impact often lies in areas like productivity, which are harder to quantify. Many companies are already promoting its democratization, aiming to create an effect similar to what digitalization achieved. The key is to accurately calculate the ROI of Generative AI use cases to support serious investment.

Q: Could you share a specific example to help us get a clearer picture?

JMA: We work on intelligent agents, AI entities that communicate with each other and with humans through natural language. These agents come together to automate a business process, interacting with applications and databases to make collective decisions and carry out actions directly. They’re more versatile and sophisticated than traditional process automation systems. Before long, we’ll be interacting with them in our daily lives without being able to tell them apart from humans.

Q: Will such technology be accessible to all companies?

JMA: Yes. The office tools we use already have Generative AI features, like helping us write emails faster or summarize meetings, which creates the impression that AI is just another work tool. The challenge is in implementing Generative AI in critical processes with the same rigor as traditional AI, ensuring it’s ethical, secure, and reliable. Companies that combine AI democratization with a systematic approach to incorporating it into business processes will gain the most value. Those that limit themselves to simply buying LLM licenses for their employees may encounter issues with AI governance and competitiveness.

Q: How does Keepler adapt to each of these types of companies?

JMA: We identify five organizational models based on data maturity. Only a couple of these models are ideal for maximizing AI benefits—those where technology and business are more aligned, commonly known as data-driven models. We’ve supported many companies in creating a strategic plan to scale business value from their data and AI use, clearly defining the ROI of their investment, identifying new roles to establish, and selecting the optimal data technology to become data-driven. Many companies are just beginning their journey into the world of data and AI, and we’re here to help them along the way.

Q: In what way?

JMA: By activating eight accelerators to drive value from data. The first is analyzing the potential ROI of use cases. The second is optimizing companies’ access to their own data by eliminating silos. Third, it’s ensuring data quality, removing bias, and making sure it’s legally usable. Fourth is modernizing data technology. The fifth is automating the data and AI lifecycle in a way that’s observable and replicable, so employees can use AI independently without relying on experts. The sixth is recognizing that AI use cases, like any digital product, have a lifecycle and need ongoing evolution. The seventh is continually measuring and communicating the business value generated. The last and most challenging is the organizational transition, aligning business and technology as closely as possible.

Q: And regarding regulation, what can we expect?

JMA: Some people say that regulation hinders innovation, but there was a need to align AI development with, for example, personal data protection laws, especially after scandals like the one involving a well-known social network that admitted to using client data to train AI models by default, unless the user explicitly opted out. The trend is being set once again by the European Union, and other countries will gradually follow our lead.