Establishing AI Fluency

Yesterday, I led a one-hour training for a few dozen VC-backed startup founders, CTOs, and execs on a topic I’ve been thinking about a lot lately: How to build AI-fluent organizations. Not just startups with shiny AI products; but companies where AI lives in the DNA of how teams operate. You can see the slides here.

Getting teams up to speed with shared rates of AI adoption might be the greatest challenge for the modern-day business leader or CEO. When I started working with Tech:NYC on Decoded Futures last summer, I was excited about the idea of exploring how nonprofits and tech companies alike are beginning to embed AI into their organizations.

But one thing from our research surprised me:

Even the most AI-forward organizations were not AI-fluent.

As it turns out, there's a big difference between building with AI and thinking with AI. Many of the orgs we spoke with had AI baked into their product, but hadn’t begun embedding it into everyday team workflows, operations, or culture. That’s a very different type of transformation.

Here are a few ways to think about AI fluency:

The Small Team Advantage

As my regular readers will know, I've spent the past year augmenting most parts of my personal and professional workflows with AI-enabled solutions. I did this largely out of necessity. As a solo operator, I'm constantly bandwidth-constrained, time-constrained, and budget-constrained. Getting good at AI is not a mere casual curiosity for me; it's a survival tactic.

As a one-person team, in many ways, I have it easy. I don't have to explain my rapid pivots, I don't have legacy datasets weighing me down, and I don't have to upskill anyone else around me. As a result, I can operate in a surprisingly agile and flexible state.

I can dedicate 1-2 hours a week to tinkering with the ChatGPT's latest image capabilities, and then build it into my product the following week. The speed from breakthrough to tinkering to deployment is faster than ever.

One of the tenets I've been holding myself to is to learn and teach something new about AI every week. You might think this is a really aggressive schedule. But it feels to be the bare minimum given the rate of change right now. And since teaching is such an effective learning framework, I'm finding that it's forcing me to crystallize my new knowledge must faster.

So much so, that I now introduce myself like this in my decks:

Technology is Easy, Training is Harder

People love to say that building the app is the easy part—distribution is where it gets tricky. I think the same holds true for AI adoption. Getting access to the technology is easy; leveling up an entire organization collectively is much harder. And it's (still) not as intuitive as you might think.

Last year, while mentoring early-stage startups at an accelerator, I found myself recommending AI workflows to founders pitching me their own AI products. They had built AI into the product, but not into the company.

The more I began speaking with even larger organizations and institutions, even the ones publishing the most cutting-edge research on AI, the more I realized the problem was even worse. The bigger the org, the slower the shift. Things like different learning styles, unclear ownership, lack of internal expertise, discomfort with new tools, legacy systems, or strict privacy requirements all got in the way.

This is when I first started to see it clearly: AI isn't just a tool, it's a mindset.

Adopting new mindsets takes time and deliberate practice. Even as a highly-motivated, self-directed individual, it took me six months of active daily practice to establish broad AI familiarity, and about a year before I started thinking of myself as AI native.

Now imagine trying to do that across a team of ten. Or a hundred. Or thousands. Even just teaching one additional person my workflow slows me down. Doing that across departments or roles introduces a whole new layer of work. That's why what we're seeing is the single-largest collective upskilling of the workforce since the advent of the Internet.

I'm committed to "teaching back" what I'm learning from the front lines of this work, as I continue to build. So if your team is starting to explore AI adoption and you're looking for practical, hands-on ways to make it stick—let’s talk.