The AI Helper I Didn’t Expect

I never expected to become an AI enthusiast. As someone who has spent 25 years leading product and technology teams, I have seen every wave of revolutionary technology roll through the corporate world. Most of them crest and crash without changing much. So when AI assistants started trending, I was skeptical. Another tool that promises the world and delivers disappointment, I thought.

I was wrong. But not in the way you might expect.

The Helper I Did not Know I Needed

A few weeks ago, Nick Zervoudis mentioned he had set up a second brain using Claude Code. What caught my attention was not the technology itself, but how he described using it. Not as a coder’s tool. Not as an engineering assistant. As a thinking partner. A PA. A helper.

The idea stayed with me. I had been drowning in a familiar problem: too much context-switching, too many balls in the air, too many threads to track across work and personal life, and paralysis as a result. So I tried it. My setup runs on OpenClaw (an autonomous agent platform I self host on an old laptop) and Claude Code (the ubiquitous agent, which I run on both my Work and Home Macs), and it has become something I genuinely do not know how I lived without.

Here is what surprised me most: despite being based on so-called coding agents, I never write code with it.

Not Engineering — Something Much More Interesting

Let me be specific about what I mean. When people talk about AI in the workplace, they usually mean one of two things: either developers using AI to write code faster, or some abstract vision of automation replacing jobs. My experience has been neither.

Instead, I have found an extraordinarily capable personal assistant. It helps me:

• Synthesise information from dozens of sources into actionable insights
• Draft communications — from blog posts to internal strategy documents
• Write and track plans for everything: work transformation initiatives and short holidays
• Manage context across multiple projects and priorities
• Think through decisions by playing devil’s advocate on my ideas
• Prepare for meetings by summarising background and suggesting angles

None of this is engineering or coding. It is management. It is leadership. It is the actual work that fills my days.

The Cerebro Effect

I called my second brain Cerebro — Spanish for brain, yes, but also the name of a Futurama-esque character who may literally be a giant brain. There is something appealing about that sci-fi resonance: a loyal, hyper-intelligent companion who happens to be artificial but is unmistakably on your side.

My Cerebro adapts to my contexts. It knows when I am in work mode versus personal mode. It understands my writing style, my preferences, my priorities. It is not a generic AI answering to everyone — it is mine. Personalised. Tailored. Effective.

The Key Insight: I Remain Fully in Control

What makes this work — really work — is that I am not handing over responsibility. I am not delegating decisions to an algorithm. I am using AI as a force multiplier for my own judgment.

The helper does not decide. It suggests. It prepares. It drafts. I approve. I steer. I control.

This is a crucial distinction that gets lost in the AI debate. The question is not AI or humans? The question is How do humans stay in command while AI does the heavy lifting?

For me, the answer has been: have a helper, not a replacement.

What This Means for Leaders

If you are a leader reading this — CTO, CPO, VP of Product, anyone responsible for teams — here is what I want you to take away:

The AI revolution in the workplace is not primarily about engineering productivity. It is about amplifying your own capacity to think, decide, and lead. It is about having help with the cognitive load that drowns most executives.

I have experienced this firsthand. My AI helper has made me more effective at my job — not because it does things I could not do, but because it does things I do not have time to do. It frees me to focus on what matters: judgment, vision, and people.

Actionable Takeaways

1. Try it as a PA first, not an engineer — You do not need to be a developer to benefit from AI. Start with management and writing tasks.
2. Customise your setup — Generic AI is okay. Personalised AI that knows your context is transformative.
3. Stay in command — Use AI as a helper, not a replacement. Your judgment is the differentiator.