Today’s businesses operate at a relentless pace. Driven by the need for rapid results, teams often compromise on core principles. Data modelling, traditionally essential to data systems, is now increasingly being sidelined.
However, it wasn’t always like this. Data modelling used to be the first step of any serious data initiative. Teams would map out relationships, design schemas, and align technical design with business meaning.
But AI could help organisations bring the practice back into focus. How exactly? Let’s take a closer look.
The Downside of Skipping Data Modelling
Sunil Patil, enterprise data management practice lead at Genpact, told AIM, “Once upon a time, data modelling was as central to data systems as a daily stretch is to a yogi.”
“Amid hackathons, agile sprints and real-time everything, one foundational practice quietly slipped off the priority list,” he added.
“Over time, data modelling became a ‘lost art’ in the eyes of many engineers,” Patil said.
So, what pushed data modelling off the priority list? A combination of new paradigms and mounting pressure. Patil believes the real culprit is the “move fast and break things” culture in software development, which has, unfortunately, infiltrated the realm of data management. Schema-on-read promised flexibility, and such agile methods left little room for deliberate design.
Patil highlighted that data engineers were no longer shaping data; they were simply moving it around. Their role devolved from modelling data to managing pipelines. At the same time, data volumes exploded and formats diversified, structured tables now sit alongside images, documents, sensor streams, and free-text logs. Modelling that chaos can feel like trying to map a rainforest as it grows.
The human gap made things worse. The bridge between business logic and technical design, the classic data modeller, has all but disappeared. “Business stakeholders speak one language, technical teams another, and the translator—the classic data modeller fluent in both—is often missing,” he added.
Ineffective data modelling techniques often cause businesses to underutilise the potential of the data available. This deficiency results in the loss of vital information and context.
Agentic AI to Bring Back Data Modelling
Patil suggested that if clean, structured data is a prerequisite for AI, why not use AI to obtain it? Enter Agentic AI—intelligent co-pilots that can help automate the once-manual craft of data modelling.
He explained that Agentic AI refers to autonomous systems that operate within the data stack, continuously refining, organising, and enriching models with minimal supervision.
Emphasising the usefulness of AI, he said, “In the context of data modelling, that means AI that can understand your data, propose schema structures, and even implement changes—all while you focus on the big picture.”
The new generation of AI agents can bridge the language divide between business and IT, enforcing standards and translating requirements into machine-readable design.
Moreover, the payoff is more than just speed. By reintroducing structure into the data lifecycle, teams reduce rework, improve governance, and finally begin delivering AI that works in practice, not just in proof of concepts.
He added that agents are autonomous; they don’t rely on human intuition to interpret data. The data must make sense to them.
The Future With AI Can Turn Things Around Again
The future of data isn’t just big; it’s smart. And that future needs context as much as it needs compute. Patil highlighted that, for years, organisations tried to run before tying their shoelaces. Now, the stumble is evident in the form of fragmented pipelines, low trust in data, and AI projects that never quite land.
Bringing data modelling back doesn’t mean slowing down. With Agentic AI, it means moving faster without compromising on clarity or control. In a world where governance, speed, and AI-readiness all matter, modelling is no longer a nice-to-have; it’s the foundation that will help grow further.
Organisations that win will be those that fuse human insight with intelligent automation. And in doing so, they’ll rediscover the old discipline of modelling, not as a forgotten art, but as a strategic necessity.
