Demystifying AI: Real leaders, real impact

How are real businesses using AI today? Not in theory. Not in sandbox pilots. But in the actual mess of legacy systems, regulation and operational complexity?

That was the focus of Session Two: Demystifying AI, a panel that brought together transformation leaders from Merlin, Howden, S&S and Vega IT to share what it really takes to embed AI in the enterprise.

Nataly Matic, Head of Consulting at Sullivan & Stanley, kicked off the session with a quiet moment that cut through the noise. She introduced us to Dave, who she told us runs logistics for a mid-sized manufacturer in the Midlands. One sweltering Tuesday morning, he finds his team drowning in late orders, stock errors and supplier chaos. Instead of calling a crisis meeting, he does something different. He opens ChatGPT — a tool that still feels like a “dirty little secret” to many — and asks it to write a script to pull data from multiple systems, cross-check it and flag high-risk suppliers. Two hours later, he has a working tool. By the end of the week, he’s saved 17 hours of manual work, and solved most of the problem.

“That’s not moonshot theory,” Nataly said. “That’s just augmentation on a Tuesday morning.”

It was a powerful reminder: AI isn’t just for Silicon Valley. It’s for real people solving real problems in the real world. The panel that followed carried that same energy forward, offering unfiltered insights on what it really takes to embed AI into your organisation.

The view from the frontline: Merlin

First up was Kinnari Ladha, Chief Data Officer at Merlin, who shared a candid view of how her team is navigating the gap between AI potential and organisational reality.

“AI doesn’t work in isolation. It needs data, it needs people and it needs a business model that can flex,” Kinnari said.“We’re not trying to boil the ocean. We’re starting with solvable problems and building trust from there.”

For Merlin, that has meant focusing on specific pain points where AI can drive measurable impact — such as unstructured data processing, policy administration and frontline automation.

“One thing we learned? Don’t get distracted by hype. Ask: what are we already doing today that AI could make better?”

She also acknowledged the cultural shift required to embed AI responsibly.

“There’s a real difference between buy-in and belief,” she noted. “It’s easy to get executives excited in the boardroom. But it’s harder to get teams to trust what the algorithm is saying when it lands on their desk.”

Navigating risk and opportunity: Howden

Paul Hillier, Managing Director, xTrade at Howden, offered the insurer’s perspective - one shaped by a heavily regulated environment, complex data estates and intense competitive pressure.

“For us, AI isn’t about replacing people. It’s about removing friction from workflows so our people can focus on what matters,” he said.

Howden is already seeing early success using AI for document classification and automated orchestration across systems, especially in areas like contract review and claims triage with massive time reductions in some processes.

Paul also addressed the common misconception that AI is a one-and-done investment.

“It’s not just about spinning up a model. You need governance, auditability, a training loop and people who know how to manage that lifecycle.”

His advice to peers?

“Start where there’s a clear business problem and a clear dataset. And build the narrative early - because trust is built long before the first pilot goes live.”

Engineering with intent: Vega IT

Bringing the technical implementation lens was Nenad Perčić, Partner & Director, Digital Strategy at Vega IT, who stressed the importance of applied AI, not innovation theatre.

“Most of our clients don’t need a chatbot. They need a way to reduce rework, automate handoffs or speed up compliance. That’s where AI can add real value.”

He highlighted Vega IT’s focus on rapid prototyping using digital twins — essentially AI-driven models of a client’s business process that let teams visualise change before it happens.

“If we can show you a working version of your new process next week, not just a spec,  then the conversation changes. It becomes: how do we make this real?”

Straight talk from the frontline: S&S

Archie Cobb, Senior AI Consultant at S&S, brought a dose of practicality to the conversation. He urged leaders to shift from top-down innovation to organisation-wide empowerment.

“I challenge organisations to move from the lab to the crowd. Everyone in the organisation needs to become just a little bit more tech-savvy,” he said.“If someone saves 20 hours with AI, don’t give them 20 more hours of admin. Reward them. That’s how you change culture.”

He also took aim at the idea that AI is only for the digitally native:

“It frustrates me when I hear, 'This is for the next generation.' It’s not. AI is everyone’s business now.”

On where to begin, Archie kept it simple:

“Don’t wait for the perfect use case. Find something messy that matters. That’s where you’ll learn the most.”

What we learned

The panel was packed with practical insights, and several key themes emerged for anyone serious about moving from AI hype to action:

• Focus beats scale: Start with one meaningful problem. Solve it. Then scale.
• Trust is everything: Whether it’s regulatory sign-off or team adoption, trust is the real currency in AI transformation.
• Execution is cross-functional: Tech, ops, data, compliance and change teams must work as one — or nothing lands.
• Show, don’t tell: Live demos, prototypes and fast feedback loops are essential to gaining traction and buy-in.
• Culture eats algorithms for breakfast: Even the best models fail if people don’t believe in them — or if processes don’t flex to accommodate them.

What made this session stand out wasn’t just the credibility of the speakers, it was their willingness to share the hard stuff: the workarounds, the blockers, the real lessons learned.

“We’re not in the business of AI. We’re in the business of change,” Kinnari said at one point. “AI just happens to be one of the most powerful tools we’ve ever had.”

It was a sentiment that resonated across the room,  and across industries.

In our next blog, we’ll unpack Danilo McGarry’s keynote, where he laid out the global AI trends shaping the next five years and why “you’re not just competing with your competitors anymore. You’re competing with exponential.” Stay tuned for Part 3.