How to avoid becoming an “AI-first” company with zero actual employ of AI

Think back to the first time you heard that your company was falling apart First, artificial intelligence?

Perhaps it passed through all hands that were different from the others. The CEO said, “By Q3, every team should integrate AI into their core workflows,” which caused a shift in the energy in the room (or on Zoom). You saw a mixture of excitement and anxiety flow through the crowd.

Perhaps you were one of the curious ones. Perhaps you already have a Python script that summarizes customer feedback, saving your team three hours a week. Or maybe you stayed up slow to see what would happen if you combined a dataset with a gigantic language model (LLM) prompt. Maybe you are one of those who have already allowed curiosity to take you to unexpected places.

But this announcement felt different because suddenly what had been a silent act of curiosity now became a line in the corporate OKR. You may not have known it yet, but something fundamental has changed in the way innovation will happen in your company.

How innovations are created

Real transformation rarely looks like a PowerPoint version and almost never follows an organizational chart.

Think about the last time something really useful came to work. It wasn’t because of a vendor offering or a strategic initiative, right? More likely, someone stayed up slow at night when no one was looking, found something to cut down on work hours, and mentioned it the next day at lunch. “Hey, try this.” They shared it in a Slack thread and within a week, half the team was using it.

The developer who used GPT to debug the code was not trying to exert strategic influence. She just had to get home to her children early. An operations manager who automated his spreadsheet did not need permission. He just needed more sleep.

This is the concealed architecture of progress – these informal networks where curiosity flows like water through concrete… finding every crack, every opening.

But look what happens when leaders notice this. What was previously effortless and organic becomes mandatory. And what used to work because it was free suddenly stops being so effective the moment it is measured.

The massive retreat

It usually starts off calmly. Often, when a competitor announces recent AI features – such as AI-driven deployment or end-to-end support automation – they claim a 40% performance enhance.

The next morning, the CEO calls an emergency meeting. The room stops. Someone clears their throat. You can also feel everyone making mental calculations about their job security. “If they’re so far ahead of us, what does that mean for us?”

This afternoon, your business has a recent priority. Your CEO says, “We need an AI strategy. Yesterday.”

Here’s how this message typically spreads through the organizational chart:

• At the executive level: “We need an AI strategy to stay competitive.”

• At the VP level: “Every team needs an AI initiative.”

• At the managerial level: “We need a plan by Friday.”

• At your level: “I just need to find something that looks like artificial intelligence.”

Each translation increases the pressure while removing understanding. Everyone still cares, but this translation changes the intention. What starts as a question worth asking becomes a script that everyone blindly follows.

Ultimately, the execution of the innovation replaces the thing itself. There’s a strange pressure Look like you’re moving speedy, even if you’re not sure where you’re actually going.

This is repeated across industries

The competitor declared that he was focusing on artificial intelligence. Another publishes a case study on replacing support with LLM programs. The third provides a chart showing productivity growth. Within days, boardrooms around the world began repeating the same message: “We should do this. Everyone else has already done it and we can’t be left behind.”

So the work begins. Then come task forces, town halls, strategy documents and goals. Teams are asked to submit initiatives.

But if you’ve been through this before, you know there’s often a difference between companies announce and what they actually are Down. Because press releases don’t mention pilots that get stuck, teams that quietly revert to the elderly way of doing things, or even tools that get used once and then abandoned. You may know someone who was in one of these bands, or maybe you were even in one yourself.

These are not technological or intentional errors. ChatGPT works well. Teams want to automate their tasks. These failures are organizational in nature and happen when we try to imitate the results without understanding what caused them in the first place.

So when everyone is innovating, it becomes almost impossible to tell who is actually doing it.

Two types of leaders

You’ve probably seen both, and it’s very effortless to tell which type you’re working with.

The entire weekend is spent prototyping. They try something recent, fail halfway through, and still show up on Monday saying, “I built this with Claude. It crashed after two hours, but I learned a lot. Want to see? It’s very simple, but it might solve the problem we were talking about.”

They try to build understanding. You could say they actually spent time with the AI ​​and struggled with hints and hallucinations. Instead of trying to sound confident, they talk about what went wrong, what almost worked, and what they’re still figuring out. They invite try something recent because you think there is still room to learn. This is what participatory leadership looks like.

The second sends you a directive in Slack: “Leadership wants every team to use AI by the end of the quarter. Plans should be made by Friday.” They enforce compliance with a decision that has already been made. You can even hear it in their language and how confident they sound.

A curious leader is gaining momentum. Performative builds resentment.

What actually works

You probably don’t need someone to tell you where the AI ​​is running. You already know because you saw it.

• Customer service: LLMs really assist with Tier 1 tickets. They understand intent, design basic answers and complexity of routes. Not perfect, of course – I’m sure you’ve seen failures before – but good enough for it to matter.

• Code assist: At 2 a.m., when you’re half-delirious and your AI assistant suggests exactly what you need, you feel like you have an overcaffeinated junior developer who never grades your forgotten semicolons. You save first minutes, then hours, and finally days.

These diminutive, cumulative victories add up over time. These aren’t the impressive transformations promised in decks, but the kind of upgrades you can rely on.

But beyond these zones, the situation becomes unclear. Revolutions based on artificial intelligence? Fully automated forecasting? You watched these demonstrations and noticed that enthusiasm waned after the pilot began.

Have the creators of these AI tools failed? Hardly. Technology is evolving, and products built on it are still learning to walk.

So how can you check whether implementing artificial intelligence in your company is realistic? Straightforward. Just ask someone in finance or operations. Ask what AI tools they employ every day. You might get a brief pause or an apologetic smile. “Honestly? Just ChatGPT.” That’s all. It’s not about a $50,000 enterprise-class platform. dollars from last quarter’s demo or an steep software package on board. Just a browser tab, the same as any student writing an essay.

You can make this confession yourself. Despite all the mandates and initiatives, your most powerful AI tool is probably the same one everyone else uses. What does this tell us about the gap between what we should be doing and what we are actually doing?

How to make changes in your company

You’ve probably discovered this yourself, even if no one has ever put it into words:

1. Model what you mean: Remember that engineering director who shared a screenshot of her messy live coding session with Cursor? You learned more by watching it debug in real time than from any polished presentation, because vulnerabilities go further than directives.

2. Listen to the edge: You know who is actually using AI effectively in your organization, and it’s not always those with the word “AI” in their title. They are curious people who quietly experiment, finding what works through trial and error. And this knowledge is worth more than any analyst report.

3. Create permission (not pressure): People willing to experiment will always find a way, and the rest cannot be forced. The best thing you can do is make the curious feel protected and stay curious.

We live in this strange moment, caught between the AI ​​promised by vendors and the AI ​​that actually exists on our screens, and it’s deeply uncomfortable. The gap between product and promise is gigantic.

But from this discomfort, I learned that the companies that will thrive are not those that were first to adopt AI, but those that learned through trial and error. They stayed in this discomfort long enough for him to teach them something.

Where will you be in six months?

By then, your company’s AI mandate will have brought to life department initiatives, vendor agreements, and maybe even recent hires with the word “AI” in their title. Dashboards will be green and the dashboard deck will have an entire slide dedicated to AI.

But what will change significantly in the silent spaces where your work takes place?

Maybe you’ll be like the bands that never stopped their silent experimentation. Your customer feedback system can pick up on patterns that people miss. Your documentation can update itself. It is likely that if you built before the mandate, you will continue to build after it expires.

This is the concealed architecture of true progress: Patience and a complete lack of concern for performance. Doesn’t create great LinkedIn posts and resists grand narratives. But it transforms companies in a truly lasting way.

Every organization now faces the same crossroads: appear to be innovating, or create a culture that fosters true innovation.

Pressure up to make innovation is real and growing. Most of the teams will relent and join the theater. But some people understand that curiosity cannot be forced and progress cannot be achieved. Because real transformation happens when no one is watching, in the hands of people who keep experimenting and keep learning. This is where the future begins.

Siqi Chen is the co-founder and CEO of Runway.

Read more in our guest authors. You might also consider submitting your own entry! See ours guidelines here.