S02E02 – AI On the Books: Build vs. Buy vs. Borrow

In this episode of The Due Diligence Show, Adam Jaques and Luke Silcock dive headfirst into the tangled mess of decisions facing companies that want to operationalize AI. Do you rent it from a hyperscaler? Build it in-house? Run it from your lab like a mad scientist in a hoodie?

From cloud marketplaces to model routing, on-prem servers to DIY open-source stacks, this episode is a whirlwind tour of what it actually takes to integrate AI into your tech stack—without getting locked into expensive vendor choices or drowning in hype.

Key topics include:

• The strategic choices of AI adoption: buy, build, or borrow?

• How to compare models on quality vs. cost (and what “tokens” really mean)

• The role of emerging tools like model routers and agentic frameworks

• The real-world limits of tools like Microsoft Copilot

• How to avoid sprawl and vendor lock-in in an increasingly modular ecosystem

• What OpenAI + Jony Ive might mean for the future of AI hardware

It’s a practical, slightly cheeky, but very real look at the evolving landscape of enterprise AI enablement. Whether you’re a CTO, investor, or just figuring out why the AI button keeps showing up in your CRM — this one’s for you.

HEAR THE PODCAST

Find all the other episodes here.