// in this post
People ask me this a lot, usually not in these words. It comes out as "should we be doing more with AI" or "are we behind." Underneath, it's the readiness question, and I've sat across the table from enough organisations now to have a genuine answer, not the diplomatic one.
Readiness has almost nothing to do with how many AI licences you've bought. I've seen organisations with AI on every desktop and less to show for it than one where four people use it seriously. The licence count measures spend. It doesn't measure whether anything's actually going to change.
What I actually look for
Real sponsorship, not tacit permission. Someone senior enough to protect the effort has to actually own it, not just approve a budget line and step back. The difference shows up the first time AI adoption gets uncomfortable, when a workflow needs to change and the people doing it push back, or when the first attempt doesn't work. Tacit permission evaporates at that exact moment. Real sponsorship is still in the room.
One use case you can actually name. Not "AI across the business." A specific thing: this report, this reply, this decision. Organisations that start everywhere end up with a dozen shallow pilots and nothing to point to. The ones that start with one thing, done properly, have something real within a quarter, and something to build the next one on.
Willingness to change how the work happens, not just add a tool to it. This is the one people resist most, because it's the one that costs something. Buying a licence is easy. Redesigning the workflow around what the tool now makes possible, including the parts that were never about the tool, means someone has to be willing to change how a piece of work gets done. If the honest answer is "we'll use it exactly the way we already work, just faster," that's not the same as being ready.
None of these three cost money to have. That's the part people find least comfortable to hear, because it means the barrier was never really the budget.
What this looks like in practice, not in theory
It's worth seeing this done seriously rather than just described. Anthropic's own team has published a look at how they actually use Claude Code day to day: real workflows, real decisions about what to hand over and what to keep checking.
It's a useful reference point precisely because it's people doing real work rather than a demo built to impress, the same judgement calls about verification and trust that show up in any serious use of AI, at any organisation, once the novelty wears off. That's the bar. Not "we tried it," but "here's how it's actually woven into how we work."
The honest bit
I build things with AI constantly, and I've had workflows fail, output I trusted that I shouldn't have, and weeks where the tool wasn't the bottleneck, I was. Readiness was never about having it all figured out first. It was having someone senior willing to stay in the room, one thing worth doing properly, and an actual willingness to let the work change shape. Everything else follows from those three or it doesn't follow at all.
If you want a considered view of where your own organisation stands against this, that's exactly what the 5-minute diagnostic is for, a live conversation, not a form, that gives you a tailored read on where you're actually stuck. If you'd rather talk it through with a person first, get in touch.
If the question you're actually asking is about your own role rather than your organisation, that's a different piece: the AI resilience audit looks at which parts of an individual job are actually defensible against automation.
FAQ
How do I know if my organisation is ready for AI?
Three things predict it better than anything else: real sponsorship from someone senior who'll defend the effort when it gets hard, one specific use case you can name rather than 'AI everywhere', and genuine willingness to change how a piece of work actually happens, not just add a tool on top of it. Licences and enthusiasm predict almost nothing on their own.
Is buying AI licences the same as being ready for AI?
No. Licence counts measure spend, not readiness. Plenty of organisations with AI on every desktop have made no more progress than one with a handful of thoughtful users, because readiness is about sponsorship, focus and willingness to change work, not access to a tool.
What's the difference between this and the AI diagnostic?
This is one person's honest view of what readiness looks like, worth five minutes to read. The diagnostic is a live, adaptive conversation about your specific organisation that produces a tailored read-out. Read this first if you want the thinking; run the diagnostic when you want the answer for your team.
This is about my organisation. Is there something for my own role?
Yes, a different piece: the AI resilience audit looks at which parts of an individual's job are actually defensible against automation, a personal question rather than an organisational one.