Telefónica is shifting its AI focus from technology to people, as the company's digital product teams recognize that the biggest constraint is no longer what machines can do, but how quickly humans can relearn and adapt. The change demands new roles, a different culture, and a company-wide plan to avoid leaving anyone behind.
From executors to decision-makers
The shift moves IT and development professionals away from generating output and toward making decisions. AI does not dilute responsibility-it elevates it. Teams now provide context, review AI-generated work, and exercise judgment at critical decision points.
Human in the loop becomes a necessity, not an option, because AI can make mistakes, go off track, or produce more than requested. The machine proposes; the human validates. This requires what the team calls product sense, design sense, and engineering sense-judgment that is hard to codify but essential for quality.
Such a shift cannot happen team by team. It requires a roll-out plan designed for the whole organization, because the hardest part is not the tool but the cultural change. Professionals who haven't had time to relearn risk falling behind, and the company sees a responsibility to support them through the transition. For IT and development teams, building this judgment often requires targeted training, such as courses on AI for IT & Development.
Speed, discovery, and new risks
AI also accelerates innovation. Testing a hypothesis has gone from expensive and slow to fast and cheap. Teams can run continuous discovery workflows, discard what doesn't work early, and focus investment on what does. Innovation becomes a process of continuous learning rather than guesswork.
Once a decision is made, the rest of the development cycle shortens: precise specifications are written faster, code is generated more quickly, and compliance, security, and QA checks can be integrated at speed. The value lies not just in speed-it's in the time gained to think more clearly and prioritize wisely.
The risk is falling into the trap of doing more just because you can. Greater capacity without focus can waste resources and create noise. The "sense" mentioned earlier-knowing what to build and why-becomes critical again.
Generative AI and agents: autonomy under supervision
Telefónica distinguishes between generative AI, which produces content based on intent, and AI agents, which chain tasks, execute workflows, and coordinate processes. Agents approach autonomous execution, but the more capable they are, the more important it is who governs them.
"An agent without human supervision isn't autonomy; it's delegated risk," an internal post said. The key question is no longer what agents can do, but what humans decide they should do, within what context, and under what control. For teams working with agents, this means designing oversight mechanisms from the start. As AI Agents & Automation become more common, the need for clear governance and human judgment intensifies.
Why this matters for IT and Development
The transformation at Telefónica highlights a concrete shift for IT and development professionals: the value you bring is no longer measured by lines of code or output volume, but by the quality of your decisions. AI handles execution; you provide context, ask better questions, and validate results. This demands a different skill set-one centered on judgment, governance, and continuous learning.
Teams that invest in building engineering sense and learn to govern AI agents effectively will be positioned to lead, while those that treat AI as just another tool risk being outpaced. The bottleneck is human, and the speed at which you relearn now determines the pace of your work and your career.
Your membership also unlocks:
