Why I, the CEO, am personally building our AI strategy

Last year, I did something that raised a few eyebrows on my leadership team: I made myself the de facto product manager for our AI strategy. Not because we don’t have talented product people — we do. But because I’m convinced that AI is one of those rare inflection points where the CEO needs to be hands-on, not just “aligned.”

I’ve been experimenting with AI tools since ChatGPT landed in 2023. Unlike other tech hype cycles — I remember an investor asking me about our cryptocurrency strategy back in 2021, and my answer was “none” — this one felt different from the start. Over the 2024 holidays, I went deep: prototyping with Cursor, brainstorming product ideas with ChatGPT, using voice mode to dictate rough drafts while away from my desk. What I took away was a conviction that AI was no longer a “keep an eye on it” item. It was clear then that AI was a right-now, every-department, every-person transformation.

Through 2025, AI became a core part of all my workflows. Now, AI is an organizational accelerator across the board at NetBox Labs. Personally, I’m shipping more product and technology than I have in over a decade — not because I’m working harder, but because AI tooling has fundamentally changed what one person can do.

And transformations like that can’t be managed by committee.

Why AI is too important to delegate

Here’s what I’ve learned watching strategy play out across startups and enterprises: the things that get delegated early tend to get domesticated early. They get scoped into a neat workstream, assigned a quarterly OKR and slowly lose their disruptive potential. AI doesn’t fit in a box right now and forcing it into one is a mistake.

But this goes beyond product management. I’m not just PM’ing our AI strategy — I’m directly building products and features, driving internal experimentation and pushing AI into every corner of how we operate. AI tooling and capabilities are evolving too fast, in too many dimensions, to delegate. The landscape shifts week to week: new models, new capabilities, new paradigms for how humans and AI work together.

As a CEO, you need to develop your own feel for the impact of AI — what it can and can’t do today, where it’s headed, how it changes the work — and bring those learnings across teams. You can’t get that from a briefing doc.

I sit at the intersection of product, engineering, go-to-market and operations. I see the connections that individual teams can’t — where an AI experiment in marketing could reshape how we think about our product onboarding, or where an engineering prototype could unlock a whole new customer conversation. When I’m personally involved, building with these tools myself, those dots get connected faster.

This isn’t about ego or micromanagement. It’s about pattern recognition at the speed the moment demands. The companies that will win are the ones whose leadership treats AI not as a departmental tool but as a strategic capability.

Speed now matters more than perfection

At NetBox Labs, we’ve adopted a bias toward shipping. Since 2024, we’ve shipped many AI features — some of which have gained rapid adoption, like NetBox Copilot or the NetBox MCP server, and others that haven’t resonated or were overtaken by the evolving AI landscape. That’s fine. The point is to stake the ground, learn and iterate — not sit in a conference room perfecting a roadmap that’ll be obsolete by the time you execute it. Recently, one of our customers asked us for better “best practices” content for working with NetBox’s APIs. In less than a day, we shipped a set of skills for AI agents that has seen quick interest and adoption. That kind of cycle time — customer request to shipped product in hours — is what AI enables.

This same principle applies internally. I tell my team: don’t wait until you’ve mastered a tool to start using it. Use it now, stumble, figure out what works and share what you learn. There are no “rules” for how AI fits into workflows yet. We’re all learning as we go, and the patterns will look different by this time next year.

Key takeaways: How to encourage company-wide AI adoption

If you’re a CEO or senior leader thinking about how to drive AI adoption across your organization, here’s what I’ve found works.

• Make it personal, starting at the top. I don’t just endorse AI use — I show my work. When I build a feature using Claude Code or prototype a product idea by feeding requirements into Claude, I talk about it openly. I even presented my use cases at our recent company offsite. Leaders using AI tools visibly and vocally give everyone else permission to do the same.
• Reframe the culture around AI. One thing I noticed early on is that people feel sheepish about using AI, like it’s cheating. We had to actively dismantle that mindset and frankly, are still working on dispelling it. Using a calculator isn’t cheating at math. Our job is to get stuff done, and teams that make effective use of these tools will outperform teams that don’t. I want to hear how people are using AI and how it’s helped them succeed — not whispered confessions, but demos and stories shared in team workshops.
• Lower the barrier to experimentation. If someone on your team finds a tool that could accelerate their work, don’t make them write a business case. Let them spend a few dollars and a few hours trying it. You can’t be wasteful and you can’t get distracted, but don’t be shy about trying new things. The cost of a missed opportunity dwarfs the cost of a failed experiment.
• Show people where to start. “I’m just not sure how to get started” is the most common thing I hear. So I share concrete examples from my own workflow: using agents to research and brainstorm to validate product feasibility, building detailed PRDs with Claude Code and then feeding them back in to generate implementation plans, continuing further to directly build and ship features, even analyzing data by connecting Claude Cowork with our CRM, Linear and call transcripts. These aren’t exotic use cases. They’re everyday work, done faster.
• Carve out time and space for people to experiment. At our recent company offsite, we dedicated an entire afternoon to learning from one another in-person. This helped everyone learn from “experts” within the company who could help them get started or uplevel their use case. Prioritizing making time for this experimentation demonstrated to the team that this truly is a company priority — not a fad.
• Connect it to individual growth. This is something I care about deeply beyond the business case. Most of us will have careers that extend well beyond our current roles. Being AI-native is going to be a defining professional skillset. I want NetBox Labs to be a place where everyone develops that skillset — not just because it helps our company, but because it matters for their careers.

The real risk is inaction

This article is published as part of the Foundry Expert Contributor Network.
Want to join?