Who gets to cook the AI?

This is the third big question in my head at the moment around work, AI and being human. For the rest of the open questions check out the first article.

Open question 3 - How does the organisational ownership of AI shape its priorities, use cases, and impact on the ground? 

Have you ever tasted a black lime? 

Bitter, smoky, sour, funky.

They’re made by taking small, fresh limes and boiling them in salty water for ages before leaving them out in the sun to dehydrate. They become hard nuggets of flavour. 

Middle Eastern dishes often grind the limes into black powder and then rub it on meat before cooking to add a certain citrus funk to the flavour. 

I have a jar of black limes in my pantry. 

Alongside my Kashmiri Chili, Sichuan peppercorns, and palm sugar. 

I like food. 

I like interesting food. 

When I cook, I want to do something creative and surprising. 

That’s because I cook to relax. 

Lots of Saturdays start with me poring over a recipe and working where I’m going to have to drive to source the next exotic ingredient. 

I cook messy. I cook long. I cook to cook.  

My wife cooks to feed. 

She makes sure everyone’s going to have enough. She keeps it as cheap as possible. She keeps everything clean and fuss-free. My whole thing is fuss. 

For her, food is fuel. For me, food is art. 

Dinner in our house can be radically different, depending on who is cooking. 

Because we’ve got completely different priorities, we end up with very different “implementations” (you can see where I’m going). 

Even though, in the end, everyone eats, how we allocate time, budget, flavour, all change dramatically based on who is cooking. The chef changes the meal. 

I have a CIO friend who is a genuine AI early-adapter. Even though he leads the technology team at his enterprise level organisation, he loves to get into the weeds with the tools. Normally I’d say it was the wrong approach for a manager, but he genuinely makes it work.  

Lately he has been building lots with Claude. He’s created some cool solutions for his organisation. One thing I’ve noticed though, is that in his organisation, AI ownership has defaulted to him and his team.  

They are the ones who have designed the solutions. They manage the data that the model sits on. They look after the on-premise server. Totally reasonable – and I bet it’s probably the most common scenario in organisations – IT holds the keys to AI. 

But it has some drawbacks. 

It means that AI use cases need to fit into the IT roadmap. This wouldn’t be too much of an issue, except just like in so many organisations, there can be communication and priority breakdown between IT and the rest of the business. In fact IT and business (everywhere) have often argued about priorities long before AI came along. AI ownership just gives the business teams something else to queue for.  

There could be a genuinely useful use case that the Sales team could utilise AI for tomorrow, but they have to raise a ticket and wait for other developments. In fact, managing support tickets is one of the first use cases IT have put to AI. Because that’s their priority. 

Contrast this with a conversation I had earlier in the year with a head of sales for a Membership Organisation. He proudly shared with me that they had let their entire IT team go. “They take too long to get anything done. We have been moving so much faster with tech and AI since we made the decision.” 

His priority was action. He wanted to execute quickly and move to ROI. I raised the issue of governance with him, and guardrails. He got the concern, or he said he did, but he just wanted to get tools into his team’s hands fast.  

Different chef.  

Last month I sat in an AI readiness workshop with a bunch of business leaders. I was seated next to the head of HR in an insurance firm. During the morning tea break, she made the case to me that really, AI should be owned by the HR team. 

“AI is going to change so many of the workflows for our people. It shapes expectations and KPIs and the whole way work gets done… We need to be leading this strategically. If we leave it to another team, our people will suffer.”  

That cracked this question open for me. 

The question of ownership. 

Which Chef gets to cook? Because the chef changes the meal. 

What does “owning AI” mean in an organisation? When a technology becomes so pervasive so quickly, can anyone really “own” it?  

There must be someone. Some team writing the budget. Deciding the use case. The one team who is in every call. 

So, who should that be? IT? HR? Maybe AI should belong to Marketing?? 

Again, like so many of these open questions, I don’t have any solid answer. But I think we should call it out. Because the chef changes the meal. 

 If IT are leading the AI race in your org, it will come with a certain flavour. It will have the architecture clear, but maybe fuzzy on the actual business benefit. 

Strong on data, poor on internal comms.  

If it’s Sales holding the tongs over the BBQ, you know there’ll be quick value-add… but will security be front of mind? Is governance built in to a Sales-driven AI project, or an afterthought. 

If it’s HR then I’m sure that the human system is being thought through, but is there enough rigour in the HOW of implementation. Are functions where value could be realised quickly being included or held out.  

Maybe there’s an answer that holds the tension. Some kind of Centre of Excellence or steering committee that includes a new function – CAI? 

But also… too many chefs. 

Worth having the conversation. I’ll be having mine over some lamb with black limes. 

 For suggested additional insights and links, and to receive regular updates in your inbox, make sure you sign up below!