Agentic AI is changing what the cloud is for

At the Google Cloud Summit in London, discussions centered on agentic AI, a shift that has transformed the cloud's role. No longer the primary focus, the cloud now serves as the execution platform for autonomous systems. This change prompts a critical question: which tasks should be handled by humans and which by autonomous software?

Yet, many European businesses still view AI as a mere productivity enhancement. This narrow perspective overlooks the broader opportunity to redesign operational workflows. Rather than merely questioning how AI can streamline existing processes, organizations should ask: if human hours and attention are no longer limitations, would those workflows be designed differently?

The current rush to deploy AI agents across enterprises is understandable but flawed. It's driven by the assumption that every existing workflow warrants an AI agent. In reality, the first question should be: why does this process exist at all? Many business processes were designed around constraints that no longer apply. These constraints often stem from human capacity limits, organizational silos, manual approvals, and legacy technology.

Implementing an AI agent in such a scenario may increase efficiency, but it doesn't fundamentally improve the business. Instead, it merely produces a faster version of something that should be eliminated. The goal is to identify areas where complexity can be removed and where the operating model should be restructured, rather than sustaining outdated practices with more expensive tools.

Europe's approach to AI spending is another notable point. At the summit, it was revealed that Europe allocates only about 8% of global AI investments. This disparity can be attributed to a mindset issue. European businesses often start with the technology itself, investing in platforms and then searching for problems to solve. Conversely, successful companies begin with the business problem, determining whether AI is the appropriate solution.

This technology-first approach limits potential outcomes, especially when the primary goal is cost reduction. A case in point involves two financial institutions. One European bank concentrated on cost-cutting to support its AI initiatives, while a US bank leveraged the technology to create new revenue streams. Both utilized the same technology but achieved different results due to their strategic ambitions.

The US bank focused on generating revenue by minimizing manual touchpoints, transforming a costly expense into an active revenue driver. Both institutions had access to identical technology; the disparity lay in their strategic objectives. European businesses should stop using legacy systems as a justification for stagnation, particularly in regulated industries like FinTech.

While legacy infrastructure poses real integration challenges, it should not be a reason for inaction. Established organizations possess core databases that may be difficult to replace overnight. However, this does not necessitate incorporating the same old constraints into new operational models. Critical legacy technology can be secured while building new, autonomous operating models around them.

By acknowledging the existing legacy technology, organizations can concentrate their efforts on redesigning specific touchpoints that offer the highest business value. Moreover, instead of automating existing workflows, organizations should aim to identify where complexity can be eliminated. Simplifying decisions and improving client outcomes will follow.

The cloud has always provided the foundational computational infrastructure; agentic AI redefines its purpose. Viewing this transition as a minor automation project will only enable you to execute the wrong strategies more efficiently. Redesigning the work itself is what fundamentally changes the business, and if you fail to do so, a competitor undoubtedly will.