Data and AI leadership is often portrayed as a race for the newest tools or the most advanced models. But for Kevin Cassar, Chief Data & AI Officer at TalkTalk, sustainable impact comes from something far less glamorous: strong foundations, clear value, and people who trust the work.
In his conversation with Dr. Raoul-Gabriel Urma on Data & AI Mastery, Kevin offers a rare perspective shaped by his own journey: starting as a data scientist, completing a Level 7 Data Science & AI apprenticeship, and progressing into the C-suite. That experience has given him a grounded, pragmatic view of what actually works when organisations try to scale data and AI.
One of Kevin’s central messages is deceptively simple: if you don’t get the foundations right, everything that follows becomes harder.
Foundations aren’t just about infrastructure. They include:
Many organisations rush toward advanced use cases before addressing these basics, only to find themselves slowed by rework, mistrust, or mounting technical debt. Kevin argues that investing early in foundations isn’t a delay, it’s an accelerator.
As AI becomes a board-level priority, Kevin believes data leaders must shift how they communicate.
Technical excellence alone isn’t enough. To secure buy-in, leaders need to frame AI initiatives in terms the C-suite cares about: outcomes, risk, and return on investment.
Kevin uses a simple but powerful framework when pitching AI work:
This approach helps move conversations away from models and metrics, toward impact and confidence. It also creates space for healthy challenges, particularly around data quality, explainability, and assurance.
One of the most practical insights from the episode is Kevin’s emphasis on bringing the business into delivery, not just at the start or the end, but throughout.
Rather than handing over finished models, his teams work in multidisciplinary squads, engaging stakeholders every sprint. This continuous involvement builds trust, improves decision-making, and increases adoption.
It also changes perceptions of data teams, from a service function to a true partner in delivery.
As Kevin puts it, adoption doesn’t happen at the end of a project. It’s built along the way.
AI leaders are constantly under pressure to deliver quickly. But speed comes with trade-offs.
Kevin is candid about the tension between rapid execution and long-term sustainability. Move too slowly, and opportunities are missed. Move too fast, and technical debt accumulates, slowing future progress and increasing risk.
The key, he explains, is intentional trade-offs. Leaders must decide where speed truly matters, and where it’s worth slowing down to get things right. That judgment, more than any tool, defines effective leadership in data and AI.
One of the episode’s standout moments is Kevin’s case study on using AI to triage healthcare claims.
By applying machine learning to prioritise cases, the team improved efficiency and outcomes, but only after addressing foundational questions around data quality, governance, and trust.
The result wasn’t just faster processing, but better experiences for customers and clearer decision-making for the business. It’s a powerful illustration of Kevin’s broader point: AI delivers value when it’s grounded in real problems and trusted by the people who use it.
Kevin’s own apprenticeship journey underpins his belief that learning never stops, especially in AI.
With technologies evolving at breakneck speed, static knowledge quickly becomes obsolete. What matters instead is the ability to learn, adapt, and apply new ideas responsibly.
That mindset applies at every level, from engineers to executives. Leaders who stop learning risk losing credibility and momentum.
Perhaps what makes Kevin’s perspective so compelling is his blend of technical depth and executive responsibility.
His journey from practitioner to C-suite allows him to translate between worlds, helping boards understand AI, and helping teams understand business priorities. That translation, he argues, is one of the most valuable skills a data leader can develop.
AI transformation doesn’t succeed because of hype, tools, or titles. It succeeds because leaders invest in foundations, trust, and people, while keeping a relentless focus on value.
Kevin Cassar’s story is a reminder that the most impactful data and AI leaders aren’t chasing the next trend, they’re building organisations that can learn, adapt, and deliver over time.
At Cambridge Spark, we help organisations build exactly these capabilities — combining technical expertise with leadership development to turn data and AI ambition into measurable impact.
👉 Explore how we support data and AI leaders.