As the data landscape undergoes a profound transformation, leading organisations are embracing new paradigms to harness the full potential of artificial intelligence.

Paul Wilson, data guru and General Manager of Data, Insights and AI here at NRI, recently featured on a conference panel hosted by ITNews. He shared his wisdom and practice in this shape shifting space, and how GenAI can and should be used to unlock insights from places we never thought to look.

Here’s how it went.

Why unstructured data is having a moment

Most of us are used to working with structured data, like spreadsheets and databases. But the GenAI wave is changing the game. Suddenly, unstructured data (think: emails, images, audio, social posts) is the star of the show. It is where real human context lives, and where GenAI thrives.

Paul says, “It’s like we’ve gone from neat filing cabinets to a massive, sprawling, dynamic library of conversations, images, and ideas.” And the key isn’t to tame that mess, but to set up the right environment for GenAI to work its magic.

“Unstructured data is where the real-world complexity lives, customer sentiments, emerging trends, unfiltered feedback. GenAI can sift through it to generate insights, predictions, or even entirely new content, like we’re seeing with language models and image synthesis.”

At NRI, that means aggregating data – pulling in diverse sources like text dumps, web content, and user interactions without over-sanitising them. Then making it accessible through indexing or tagging it in ways that let the AI navigate it efficiently, even if it’s not perfectly structured.

“The key is not to over polish, but you still need to exercise quality control, and clean up noise like duplicates or irrelevant outliers, but keep the authentic variability that fuels creativity in GenAI.”

Modernising without the mayhem

Data modernisation and migration are at the heart of so many IT transformations right now, and it’s a complex beast to tackle. Modernising data platforms isn’t a one-size-fits-all job. NRI has taken a blended approach—centralising where it matters, but keeping federated access to live, in-the-moment data where speed is key.

“We’ve had to prioritise flexibility over perfection. For instance, instead of forcing every dataset into a pristine, standardised structure upfront, we’ve built pipelines that can ingest and process on the fly, letting the AI handle some of the harmonisation. It’s not always pretty, but it keeps things moving fast.”

The biggest hurdle? Change management. Old habits around data collection and definition can be tough to shift. That’s why NRI pairs smart automation tools with practical training, helping to ease the burden, but also showing teams how good data practices lead to better AI outcomes.

“My biggest advice here is not to underestimate the human side. Tech can solve a lot, but if your team isn’t bought in, you’re fighting an uphill battle.

Showcasing small wins is enough, like a cool AI insight from newly migrated data, and it will go a long way to build momentum.”

What’s going to drive the next leap?

Looking ahead, NRI is focused on three key ingredients to keep pushing the boundaries of what data can do:

• Scalability: With data growing at breakneck speed, infrastructure needs to scale effortlessly. Think distributed systems, cloud elasticity, or optimised storage – don’t let data ambitions stall because you can’t process or store what’s coming in.
• Adaptability: The data landscape is always changing. What’s unstructured today might need to blend with structured streams tomorrow. It is critical to invest in tools and processes that have the flexibility to keep up.
• Real-Time Richness: The freshest insights come from live, dynamic data, like social feeds, sensors, or breaking research. The key is to build systems that can ingest and process this data in real-time.

When these three elements come together, the possibilities really open up. As Paul puts it: “We’re not just collecting data. We’re building a living engine for discovery.”

NRI’s offerings are about empowering organisations to harness the full potential of their data and artificial intelligence. For this, we focus on three core pillars that are designed to work together and independently.