Most businesses approach AI the wrong way around. They start with a tool, a vendor pitch, a proof of concept, a shiny API, and then ask what problem it solves. The result is a patchwork of disconnected capabilities, a growing list of AI subscriptions nobody is fully using, and a leadership team no clearer on where AI actually fits in their business model.
The fundamental mistake is treating AI as a technology decision rather than a strategic one.
The industrial revolution parallel
We are living through a transformation that rivals the industrial revolution in its scope. Just as steam power didn't simply accelerate existing manual processes but fundamentally restructured how industries organised themselves, what they made, how they made it, who did what, AI is reshaping the operating model of the modern business.
The businesses that thrived in the industrial revolution weren't the ones that simply bolted machinery onto existing workflows. They were the ones that reimagined what their business could be in a world where the constraints of manual labour no longer applied.
The same logic holds today. The question is not "how do we add AI to what we already do?" It is "given what AI makes possible, what should we become?"
Capability-first, technology-second
The right starting point is a clear capability model: what your business needs to be able to do well in order to deliver its strategy. From there, you can assess which of those capabilities AI can materially improve, and how.
This is the lens Oxygen Bubbles applies. We work with leadership teams to define what good looks like across their organisation, identify where AI creates real leverage, and build a roadmap that technology then serves.
This approach produces something far more valuable than a set of AI tools: it produces a coherent AI strategy that the whole organisation can orient around.
What this looks like in practice
For one financial services client, the starting point wasn't "should we use AI?", it was a capability gap analysis that revealed three areas where processing speed and consistency were limiting growth. AI solutions were then selected and sequenced to address those specific gaps, with governance structures designed from day one.
For a professional services firm, the question was about knowledge leverage: how do we ensure that the expertise of our most experienced people flows through every client engagement, not just the ones they personally touch? The AI answer to that question was very different from a generic "AI for productivity" initiative.
In both cases, the technology followed the strategy. That is the right order.
We explore the governance dimension of this in AI Governance in Financial Services, and the practical delivery model in The 8-Day AI Sprint.
If you're ready to build an AI strategy that leads technology rather than chasing it, our Breathe discovery sprint is designed to do exactly that in 5–10 working days.