In this episode of The Data Source Podcast, I’m joined by John Murphy, Co-Founder of Ailíniú, to explore what it really means to build with AI in practice.

John spent years as a Chief Architect in some of Ireland’s largest organisations before moving back into building products himself, with AI now a core part of how he works.

We discuss the difference between architecture and shipping, why AI can help teams build faster, and why that also makes it easier to build the wrong thing quickly.

John shares a practical view of how he uses AI day to day - as a thinking partner, a second brain, and a coding assistant - while also explaining where deep human judgement still matters most.

We also get into the reality of taking AI into production: deterministic answers, evaluation, observability, model changes, fallback paths, human oversight, vendor risk and governance.

For regulated firms in banking and insurance, John explains why AI governance cannot be compliance theatre. The firms that get this right are not slowing themselves down - they are creating the guardrails that allow them to move faster with confidence.

This is a conversation about AI, architecture, regulation and what it takes to ship useful, controlled and trustworthy systems in the real world.

Enjoy!

In this episode of The Data Source Podcast, I’m joined by Joe Dunleavy, European CTO and Global Head of Dava.X AI at Endava, to explore how enterprise organisations can use AI to transform how software is built and delivered at scale.

We discuss why context is king when using AI agents to produce code, and why the quality of the output depends heavily on the quality of the context provided before the agent ever starts coding. In an enterprise environment, that means more than a simple prompt - it means detailed instructions, relevant data, business context, architectural considerations, coding standards, regulatory obligations, security requirements, policy-as-code and a clear understanding of the required outcome.

Joe also shares his perspective on Flow, the methodology he helped build at Endava, and how it helps large organisations connect strategy, teams and execution so they can improve delivery and make better decisions at scale.

We also explore why the AI narrative needs to be bigger than efficiency. AI will help organisations work faster, but the real opportunity is not doing the same work cheaper - it is creating more value through better products, better decisions, better customer experiences and entirely new ways to create value.

This is a conversation about AI-assisted software delivery, context, transformation and what it really takes for large organisations to move from AI ambition to real-world impact. Enjoy!

More information about Flow, Endava’s AI-native engagement system:
https://www.endava.com/dava-flow

Joe's recent keynote at Miro Canvas conference in San Francisco:
https://www.youtube.com/watch?v=t6CLJKL8omU

In this episode of The Data Source, I’m joined by Siobhan Fairman, VP of Engineering at Compliance & Risks, to explore how AI systems and MCP are reshaping modern SaaS architecture.

We discuss what becoming AI-first really means beyond adding features, how teams are separating models, context and tools in practice, and why workflow and orchestration are becoming critical design decisions.

We also examine the human side - why this shift isn’t just technical but organisational, and why, without alignment across leadership and teams, AI-first initiatives struggle to succeed.

This is a conversation about building AI for production - with control, clarity and long-term thinking. Enjoy!

In this episode of The Data Source, I’m joined by Olivier Mertens, who leads AI Architecture at Workhuman, to unpack how AI is changing the reality of modern architecture roles.

We explore how solution and enterprise architects are adapting as AI becomes part of the core platform, why the AI Architect is emerging as a critical role, and what it really takes to move AI from experimentation into secure, scalable production.

This is a grounded discussion on designing for ambiguity, managing risk and governance, and enabling teams to use AI in ways that deliver genuine business value. Enjoy!

In this episode of The Data Source Podcast, James Bowles is joined by Richard Skinner, CEO of Phased AI. With years of experience helping companies move from proof-of-concept to production-ready AI, Richard offers a refreshing take on what it really takes to build trusted AI systems that deliver value.

Together, they unpack why so many AI projects fail to launch, the importance of deliberate testing over breakneck speed, and why humans remain essential in the AI loop. Richard shares real-world insights on quality assurance, risk assessment, and building confidence in AI tools through rigorous evaluation and user feedback.

If you're a tech leader looking to get AI right, not just live, this episode is packed with practical advice you won’t want to miss.

In this episode of The Data Source Podcast, James Bowles is joined by Patrick Callinan, Chief Data Officer at CleverCards and Fellow of the Analytics Institute of Ireland. With a background in economics, analytics, and consulting, Patrick brings a sharp lens to the financial realities of AI adoption.

Together, they unpack the often-overlooked costs of building and scaling AI, from compute and talent to data readiness and cultural change. Patrick explains why R&D spending isn't always wasted, why many AI projects still fail, and how businesses can avoid the “innovation theatre” trap by placing the right bets. They also explore how mid-sized firms can strategically embrace AI, and what every business leader should consider before diving in.

If you're a founder, CDO, or tech leader trying to make sense of AI’s real ROI, this episode is a must-listen.

I am delighted to share a recent conversation with Joe Dunleavy, European CTO at Endava where we discuss all things Agentic AI. Joe oversees the AI practice at Endava, which is focused on the use of AI to drive real business value. Joe breaks down what Agentic AI is, its current use and its future potential, sharing valuable insights for people interested in exploring the power of Agentic AI to innovate and drive business value. Enjoy!

Agentic AI is the new wave in AI innovation. Allowing autonomous pursuits of complex goals, combining LLM power with programming precision for proactive, context-aware operation in applications like robotics, complex analysis, virtual assistants and much much more.

I recently sat down to discuss AI and the Regulation of AI with Dr. Eoin Lane, and in particular, we discussed the EU AI Act and the pros and cons of regulating such a powerful, impactful, and rapidly changing technology.

Eoin is an expert in regulatory compliance regarding AI, previously leading Bank of New York Mellon's Global Data Science and AI teams, delivering award-winning enterprise-grade AI applications. He now runs Noval Consultancy, helping companies on their AI journey and regulatory compliance.

We also discussed how to approach AI initiatives and why some projects succeed and others fail, why you need to level up with AI or risk being left behind, and why the combination of AI and human skill and innovation will be the ultimate winner. Enjoy!

Introduction to the Data Source Podcast.