For most of the last decade, the industry conversation around Information Technology (IT) has been dominated by infrastructure. Cloud migrations. Distributed systems. Data lakes. Scalable architectures. And for good reason. These investments created the backbone of modern digital businesses. But after working across telecom, financial systems, retail analytics, travel platforms, and utilities—building and scaling products from early-stage ideas to global deployments—one pattern has become impossible to ignore.
Infrastructure is no longer the differentiator. Intelligence is. The organisations that will define the next decade are not the ones with the best infrastructure stack. They are the ones that can convert that infrastructure into decision-making systems. This shift did not become obvious overnight. It revealed itself slowly, across very different industries, through repeated friction, failed assumptions, and eventual breakthroughs. Early in my career, working on airline booking platforms, the focus was clear: Drive conversion and retention. Pricing experiments, funnel optimisations, and API integrations were the tools. At one point, introducing dynamic pricing logic based on user behaviour contributed to a measurable uplift in annual bookings. But what stood out was not the feature itself—it was the decision loop behind it. We were not just building systems. We were shaping how decisions were made in real time.
Later, working on retail analytics and price optimisation platforms, the scale changed. Global rollouts, multiple markets, and enterprise customers. Infrastructure mattered deeply—but it was not enough. A re-architecture of a legacy analytics system improved uptime by 40% and reduced operational costs significantly. But the real value came when pricing decisions became faster, more consistent, and more aligned with market signals. Again, the pattern repeated. Systems that helped people decide better created more value than systems that simply processed data faster.
{{/usCountry}}Later, working on retail analytics and price optimisation platforms, the scale changed. Global rollouts, multiple markets, and enterprise customers. Infrastructure mattered deeply—but it was not enough. A re-architecture of a legacy analytics system improved uptime by 40% and reduced operational costs significantly. But the real value came when pricing decisions became faster, more consistent, and more aligned with market signals. Again, the pattern repeated. Systems that helped people decide better created more value than systems that simply processed data faster.
{{/usCountry}}The shift became even clearer while building data platforms for property professionals. Here, the challenge was not infrastructure—it was trust in data. Improving data processing speed by 50% and reducing reporting latency was important. But what unlocked adoption was clarity—clear signals, reliable insights, and decisions users could act on with confidence. Across each of these experiences, the same lesson emerged in different forms. IT is no longer about systems. It is about decisions.
Today, in more recent work leading global product initiatives, the gap between perception and reality has only widened. Organisations are rushing into AI adoption. There is pressure to embed intelligence into products, to automate decisions, to do something with Artificial Intelligence (AI). But in many cases, the foundations are not ready. Data is inconsistent. Ownership is unclear. Feedback loops are missing. Decision pathways are fragmented.
In one instance, improving coordination, ownership clarity, and decision flow reduced product delivery timelines from three months to four weeks across global teams. No new AI model. No new infrastructure. Just better decisions. This is where most AI strategies fail. They attempt to layer intelligence on top of unresolved complexity. To address this, I began structuring what I had seen repeatedly across organisations into a clearer mental model—a way to understand why some systems evolve into intelligent platforms while others remain operational tools. This became what I refer to as the Decision Intelligence Stack.
At its base is data integrity. Without clean, reliable, and consistently defined data, nothing else works. Above that is context. Data needs meaning—customer segments, behaviours, operational signals. Without context, data is noise. Then comes the decision layer. This is where most organisations focus their AI efforts. But intelligence is not just about models. It is about how insights translate into actions.
Finally, the most overlooked layer: feedback. Every decision must generate learning. Without feedback loops, systems do not improve. They stagnate. What is striking is how often organisations over-invest in the decision layer while neglecting the layers beneath and above it. That imbalance is why many AI initiatives struggle to deliver real value. The implications of this shift go beyond systems. They redefine roles.
Engineers are no longer just building features. They are shaping execution logic. Data professionals are not just analysing trends. They are defining what signals matter. Product leaders are not just managing roadmaps. They are designing decision ecosystems. This is not a small evolution. It is a fundamental change in what it means to work in IT. There is also a discipline required that is often overlooked. Not every problem needs AI.
In fact, some of the highest-impact improvements I have seen came from simplification. Removing unnecessary steps. Clarifying ownership. Improving how decisions flow through systems. AI should amplify good systems. It should not compensate for broken ones. The organisations that understand this will move faster, learn faster, and create more meaningful impact. The ones that do not will continue to invest in technology without seeing proportional returns. Looking ahead, the next decade will not reward those who build the most complex systems. It will reward those who build systems that think clearly, decide effectively, and learn continuously. Infrastructure will remain important. But intelligence will define success.
(The views expressed are personal)
This article is authored by Nohit Arora, head of product/digital & AI strategy.