Is It Time for Businesses to Create a GenAI Division?

In 2023, Seth Godin wrote:

"When a new technology arrives, it is almost always at a systemic disadvantage. If we wait until the new thing is better than the old thing, we’re taking a big risk."

I agree. I have been using GenAI code IDEs and GenAI wrappers extensively. In fact, I even built one myself. But I will admit—they often fail to meet the quality standards of today's human experts. They are kind of hit-or-miss. But their current limitations shouldn't blind us to their future potential.

That’s exactly why Seth’s warning matters so much. If we sit around waiting for GenAI to become perfect, we will miss the boat altogether. The question I’ve been chewing on is: What should we do instead?

Seth has a solid rule of thumb for this:

"When a new tech or system is 50% as effective as the old one, it’s our job to learn it and understand it. And when it hits 77%, we ought to consider creating a new division, a new product line, or a new approach that adopts it."

I think we’re already there. GenAI might not blow us away just yet, but it’s definitely crossed that 50% mark and is pushing into the 77% range in some areas. So, yeah—it might be time to start thinking about creating a dedicated AI division. But what would that even look like? And what would they actually do?

# A Quick Look Back: What History Can Teach Us

Let’s talk about how transformative technologies usually roll out. Take the web as an example.

In the early days, universities started uploading their application forms onto websites in PDF format. That was... fine, I guess. It saved some hours and may be postage for the students, but the process itself stayed exactly the same. Nothing groundbreaking.

Then the second wave hit, and parts of the process became automated. Universities didn’t completely overhaul how they worked, but things started moving faster. Application processing became quicker on the admin side, though students probably didn’t notice much of a difference yet.

Fast forward to today, and everything’s changed. Universities are offering fully online courses—not just fudging around with automation, but unlocking entirely new business models. Platforms like Coursera and EdX mean they’re enrolling students from across the globe and generating revenue streams they couldn’t have dreamed of back in the PDF-uploading days.

New tech always starts by mimicking what already exists, then streamlines things, and then—eventually—changes the game entirely. GenAI is at stage one or two right now (depending on the use case), but history tells us its potential doesn’t stop there.

# Where to Begin?

So, back to the original question: If you’re running a business, what should your shiny new AI division actually focus on?

As I quoted, Alex Kantrowitz, in Will A Chatbot Take Away My Job:

"One, it is about minimizing all possible support work they can to make room for employees to come up with new ideas. Two, it is about building systems inside companies to make sure that good ideas get to decision-makers."

Here’s how I interpret that in practical terms:

1. Automate the Boring Stuff
   Start by identifying all the repetitive tasks that drag your employee's productivity. These are the low-hanging fruit for GenAI to take over. Freeing up your best people will give them space to do what they do best—innovate and think big.

2. Let People Experiment
   Don’t just hand off GenAI tools to your employees and expect magic. Give them space (and permission) to play around with the technology. Let them figure out where it works, where it doesn’t, and where it could surprise you.

   Send them to conferences, pair them with industry thought leaders, or even hold regular brainstorming sessions. The more they experiment, the more likely they are to come up with ideas that actually move the needle.

3. Put a System in Place for Good Ideas
   The last thing you want is for all that creative energy to get stuck in limbo. Make sure your team has a clear way to pitch ideas, test prototypes, and get buy-in from decision-makers. Otherwise, those game-changing ideas will get stuck in Slack threads or Google Docs purgatory.

# Why Now?

GenAI isn’t going to wait for anyone. The sooner you get comfortable with GenAI and build a framework for using it, the sooner you’ll be ready for what comes next.

And honestly, it doesn’t have to be fancy. Start scrappy, learn as you go, and let your AI division evolve over time. Most of the world’s most transformative systems didn’t start perfect—they started as something barely functional and got better with experimentation.

So yeah, if you’re running a business, this is the moment to start building that GenAI division. Not tomorrow. Not “when the tech gets better.” Now.

# TL;DR

New tech doesn’t have to be perfect to start making a huge impact on your business. If GenAI has even 50%-77% of the potential you need, dive in. Free up your best people to innovate, give them room to experiment, and make sure their ideas don’t get stuck on the drawing board.

The AI revolution might feel messy and imperfect right now, but that’s exactly why it’s the perfect time to jump in.