the Same Hiring Mistake is About to Get More Expensive

I published a version of this in 2021. It was about hiring Agile coaches. Reading it back, the argument is identical. The job titles have changed and the day rates have roughly tripled.

Organisations are moving fast on "AI transformation leads," "Chief AI Officers," and consulting firms promising to compress a multi-year change into a 90-day sprint. Some of that money will produce results. Most of it will produce a very expensive lesson about organisational readiness that a two-hour honest conversation could have avoided.

Four things to acknowledge before you start.

One: No AI expert knows everything about AI transformation. They're human, same as you. Deep capability in some areas, real gaps in others. Worth knowing before you set expectations.

Two: Certifications tell you what someone has learned, not what they can do. I know what a cake is. I can't bake one from scratch, and I certainly can't show you how. I've watched people arrive with impressive credentials, a Scrum Master or Agile Coach cert rebranded as "Agile expertise," a prompt engineering certificate rebranded as "AI strategy," and struggle to apply any of it in a live organisational environment. Knowing and know-how are two different things, and the gap between them is where most hiring decisions go wrong.

Three: What worked elsewhere won't automatically work here. I've seen organisations bring in Agile coaches from European companies, expecting them to transfer the cultural and systems context from there into a New Zealand environment. It rarely worked. The organisational conditions, the regulatory environment, the pace expectations, the relationship dynamics between teams and leaders, all of it is different. An AI expert who spent the last 5 years in a San Francisco tech company is carrying a very specific operating model in their head. That model may or may not survive contact with a NZ government agency or a regional bank that's been running for 150 years.

Four: Some AI experts are unicorns. By that I mean they only function in perfect conditions. They arrive with a clear picture of what the organisation should look like once they're done with it, and they expect the organisation to submit to that vision. When the org doesn't comply, they diagnose a culture problem or a leadership problem or a readiness problem, and they're not entirely wrong, but they're not useful either. An expert who can only work in conditions that already look like success isn't solving your problem. They're auditioning for a different organisation.

Five positions, five different problems

Low Knowledge + Low Reception = Bystander. 

Your organisation hasn't started, and isn't particularly motivated to. Bringing in an AI transformation expert here will produce a unicorn problem almost every time, because the operating conditions can't support them. There's also a structural version of this position that's worth naming separately.

In some NZ government agencies, reception has nothing to do with enthusiasm and everything to do with policy. Staff aren't empowered to make decisions on behalf of their leaders. The hierarchy and accountability structures aren't a culture problem, they're a legal and policy architecture that nobody is dismantling regardless of how good the AI use case is. 

I watched Agile coaches walk into that environment expecting holacracy and come out confused. It's not blind leading the blind, exactly. It's more that anyone who walks in thinking they can see the full picture in that context is leading people toward a very expensive dead end. Train first. Run small experiments in contained spaces. Build curiosity before you bring in someone to scale it.

Low Knowledge + High Reception = Believer. 

Let's face it, this is the best starting position. Leadership is bought in, which is the hardest thing to create. Use that cover. Invest in AI literacy across your teams first, practical workshops, hands-on tool exposure, small internal experiments that prove value in your specific context. Then bring in an experienced practitioner to help people apply what they've learned to your actual problems. In the interview, ask them to describe specifically how they'd approach your environment, not their last client's environment. A generic answer tells you everything you need to know.

High Knowledge + Low Reception = Practitioner. 

You have pockets of genuine capability, maybe a few teams doing interesting things, but no organisational pull to scale it. At BNZ, when I was there, we had exactly this: a traditional IT department running waterfall on one side, and a digital department on the other, its own reporting line, its own leadership, doing everything in an agile manner to bring retail online services to market. Both existed simultaneously. The digital teams were capable and moving fast. The org-wide reception for their approach was mixed at best. An AI coach dropped into that second group can do good work, but they'll spend half their time managing the friction at the boundary rather than coaching. The challenge in this position is political, not technical. Ask yourself whether you need executive sponsorship before you hire expertise, because without the first, you're setting up the second to fail.

Medium Knowledge + Medium Reception = Adopter. 

Stuck in the middle, probably frustrated, probably not getting the results you were promised. Adding more experts won't unstick you. Decide which axis you need to move on first, because confusing an AI knowledge problem with an AI reception problem is exactly how organisations end up cycling through consultants without changing the underlying conditions.

High Knowledge + High Reception = Native. You're in the minority. You don't need transformation expertise. You need specialists who can extend what you're already building, people with deep domain capability in specific areas, not generalists who've read the same frameworks as everyone else.

The question before the question

The matrix is a diagnostic, not an answer. The point is to prompt this conversation: 

Does the hire you're planning match your actual position, or the position you wish you were in?

Most expensive AI hires are made from the second answer.

It's worked for me, knowing the difference before I signed the paperwork.