Meeting Steve Made the AI Opportunity Feel Obvious

I recently went to the Future is Human conference in Johannesburg. It was genuinely excellent. The event was thoughtful, well run, and full of useful ideas about where AI is taking work, leadership, and business.

But I left with a slightly uncomfortable thought.

For most busy CEOs, even a brilliant AI conference is not enough.

That became clear to me after meeting Steve Holt. Steve is the kind of person who makes this whole AI moment feel very real. He is a high-performing CEO and healthcare entrepreneur, running a complex group of companies with hundreds of people and a lot of moving parts. He has the judgment that only comes from years of operating in the real world. He understands people, pressure, risk, teams, customers, and the daily reality of building companies.

That judgment is incredibly valuable.

But judgment alone is no longer enough. Not because judgment has become less important. The opposite is true. In an AI-first world, good judgment becomes even more valuable. The problem is that the tools are moving too quickly for most CEOs to translate that judgment into new systems on their own.

A CEO like Steve can attend a great conference, learn a lot, and still go back into the reality of running multiple companies, managing hundreds of people, making decisions, handling problems, and keeping the machine moving. The inspiration is useful, but it does not automatically become new workflows, internal tools, training systems, automations, dashboards, or better decisions.

That is the gap I keep thinking about.

I do not think the answer is another generic AI course. I do not think the answer is a big consulting deck. And I do not think the answer is telling every CEO to become a software engineer.

I think the answer is a person.

More specifically, I think the answer is a trusted AI right hand.

Imagine pairing a serious CEO with a young, sharp, AI-first operator. Someone who has grown up with these tools, plays with them every day, understands how quickly they are changing, and can sit close enough to the CEO to understand the real business context.

The CEO brings judgment.

The operator brings AI fluency, speed, curiosity, and the ability to build.

Together, they can do something neither could do alone.

The CEO can explain how the business actually works: where the friction is, where the risk is, which teams are overloaded, where decisions get stuck, what customers keep asking for, what reports matter, what knowledge is trapped in people's heads, and which problems are worth solving.

The AI operator can turn that into prototypes: internal assistants, knowledge bases, workflow automations, reporting systems, onboarding tools, customer support flows, training material, decision dashboards, and all the strange little pieces of operational leverage that are hard to see from the outside.

This is the idea behind Tandem.

Tandem is a way to match high-performing CEOs with AI-native operators who can become their right hand inside the business. Not as a consultant who visits once and disappears. Not as a junior intern who needs to be managed heavily. As a carefully selected, trained, and supported operator who helps the CEO convert judgment into working AI systems.

The matching is only one part of it.

The real value of Tandem is the support layer around the operator. Tandem should train these operators, give them a practical playbook, support them when they get stuck, build a knowledge base of patterns and tools, and create a community of high-performing AI operators working with serious CEOs across the country.

Every operator should make every other operator better. Every engagement should improve the playbook.

That is also why I want to build Tandem in an AI-first way from the beginning.

In fact, this post is part of that experiment. After the idea started crystallising, I asked my AI agent to start helping me build the company. It has already been helping capture the thinking, write the first drafts, update the website, organise notes in my vault, and turn messy voice notes into actual working material.

I was doing this from the beach, on my phone.

That is the point.

This is how you buy back your time and give yourself the ability to work from anywhere. It is basically like having an employee, or even a small team of employees, helping me build the business in the background while I keep moving.

That feels important. If Tandem is going to help CEOs learn how to work with AI-native people and AI agents, then Tandem itself should be built that way.

I do not yet know exactly what this becomes. It might start as a short AI opportunity sprint. It might start with one embedded operator beside one CEO. It might start with Steve, if he is interested in building this with me or becoming the first design partner.

But the core conviction is getting clearer:

The next wave of AI adoption in serious companies will not be won by the leaders who attend the most talks. It will be won by the leaders who pair their judgment with people who can build.

That is the Tandem idea.

One experienced CEO. One AI-first operator. Working side by side.