Welcome to another episode of the Data Debrief, the companion show to Driven by Data: The Podcast, where hosts Catherine Dowden-King and Kyle Winterbottom unpack Tuesday's episode, share what's been on their minds, and explore the realities of leadership, culture, and capability across the data and AI landscape.

This week, Catherine and Kyle reflect on the conversation with Justin Borgman, co-founder, CEO, and chairman of Starburst, diving deeper into why AI adoption keeps stalling at scale, the real cost of pointing powerful tools at the wrong problems, and what it means to build differentiated business capability rather than just better infrastructure.

They cover:

Why Justin's refreshingly candid starting premise — that messy, fragmented data is simply the reality most organisations are working in — cuts against the vendor instinct to promise a clean, unified solution, and why that honesty lands differently coming from a SaaS founder
The recurring pattern of businesses spending two to three years consolidating data into a single source of truth, only to arrive at the same "so what?" question — and why Starburst's founding premise of using data where it lives challenges the orthodoxy of centralisation as a prerequisite for value
The context problem that no platform solves on its own: how the same number pulled from the same source can mean two completely different things depending on interpretation, and why that ambiguity at enterprise scale can quietly corrode trust in data across an entire organisation
Justin's observation on where AI is delivering the clearest, most demonstrable value right now — coding — and what that signals for how skill sets in software development, data science, and adjacent technical roles are likely to evolve faster than most organisations are prepared for
The entry-level talent question neither businesses nor education systems have yet answered: as AI absorbs the work that once built foundational experience, where does the next generation of senior leaders come from, and who quality-assures the outputs of people who have never done the work themselves
Catherine's take on AI as fire: extraordinarily useful when understood and controlled, capable of running out of control very quickly when deployed at enterprise scale through FOMO rather than focus — and why a CFO's instinct to shut it all down is an entirely predictable response to cost spirals
Kyle's reflection on the speed problem at the heart of this AI cycle: unlike previous technological revolutions, where the pace of change gave industries time to adapt and reskill, this one is moving fast enough that many organisations and individuals haven't yet worked out what adaptation even looks like
A moment from Catherine's farming background and the latest series of Clarkson's Farm that brings the AI transition into sharp relief — precision agricultural technology that looks futuristic to most farms but is closer than people think, and what the emotional weight of replacing a working horse with a tractor tells us about how humans really respond to transformation

Kyle's thought of the week: prompted by a pattern he's been tracking across executive search processes throughout 2026, Kyle reflects on a frustrating gap between capability and communication at the senior leadership level. The people not getting the roles aren't failing on technical grounds — they're losing out on energy and enthusiasm, inability to be concise, talking around questions rather than answering them, and failure to give specific examples. Kyle's concern is that these aren't just interview problems: they're signals of how someone will perform in front of a board or a CEO, and the skills that fix them can be self-taught and improved quickly. Catherine adds a practical tip for building confidence in high-pressure communication situations using AI tools like ChatGPT or Claude as a low-stakes rehearsal partner — and shares a striking example from a full studio broadcast that shows how dramatically even experienced communicators can disappear under pressure.

This episode explores why the data and AI industry's biggest bottleneck isn't the models — it's the foundations, the focus, and the people trusted to lead the work and make the case for it.