The Myth:

We need to disrupt our industry before someone else does.

The reality: the signals about where to start are already inside your building. Your employees are the canaries. And most of them are already flying.

Half Your Workforce Is Already Doing It

A BlackFog survey published in January 2026 landed a number that makes the "should we adopt AI?" conversation feel quaint. 49% of workers at companies with more than 500 employees admit to using AI tools without employer approval. Not occasionally. 86% are using AI weekly for work tasks.

The most common use cases: technical support, sales, contracts. The stuff that keeps the business running.

But here's where it gets interesting. This isn't a rogue employee problem. It's a leadership problem hiding in plain sight. 69% of presidents and C-suite members said they're OK with unsanctioned AI use, prioritizing speed over caution. Among executives and senior managers specifically, 93% admitted to using shadow AI tools themselves.

Read that again. 93% of your executives are using unapproved AI tools at work. The same executives who haven't approved those tools for their teams.

A Cybernews survey of more than 1,000 US employees found that 59% use unapproved AI tools, and 75% of those users admitted to sharing potentially sensitive information with those tools. Customer data. Employee data. Internal documents. And 57% said their direct managers are aware of it and support it.

46% said they'd keep using these tools even if the company explicitly banned them.

This is not rebellion. This is people solving problems because the official systems don't work well enough. MIT's "GenAI Divide" report called it a "shadow AI economy." While only 40% of companies have official AI subscriptions, workers in over 90% of organizations use personal AI tools for their jobs. Nearly every employee surveyed reported using AI in some form as part of their regular workflow.

The shadow users aren't the problem. They're the signal.

Snow Melts from the Edges

Real change doesn't start with executive mandates or company-wide rollouts. It starts with 1 person solving 1 problem. That insight spreads. The tools spread. Before long, the organization starts to shift from the bottom up.

Donella Meadows described this as self-organization. The ability of a system's components to create new structures and patterns without top-down direction. The healthiest systems allow this. They create space for experimentation at the edges. They watch for signals. They amplify what works.

Vijay Govindarajan's Three Box Solution connects directly here. Box 3, creating the future, doesn't have to be a top-down initiative with a budget and a steering committee. Your employees are doing Box 3 work every time they solve a problem with an unapproved tool. They're experimenting. They're testing. They're learning what works and what doesn't. The question is whether you're capturing that learning or suppressing it.

VG's insight about Box 2, selectively forgetting the past, matters here too. The policies that ban unapproved tools, the approval processes that take months, the IT governance frameworks designed for a world where employees couldn't install their own software. Those are chains, not roots. They're past practices preventing future experimentation.

Sangeet Paul Choudary's February 2026 HBR piece reinforces this with the Shein example. Traditional fashion companies organize around the season. The executive suite decides what the market wants months in advance. Shein flipped this by reducing the unit of work to continuous small-batch experiments that test demand in real time. The learning happens at the edges, in the market, through constant contact with real customers.

The companies that figure this out don't start by looking at competitors. They start by looking inside.

The Wrong Way to Look for Disruption

Executives spend enormous amounts of time studying competitors. Attending conferences. Reading analyst reports. Hiring consultants to tell them what everyone else is doing.

Most of it is wasted.

If you spend all your time looking at what competitors are doing, you miss the signals coming from inside your own building. And if you want to look externally for opportunities, your industry or vertical is a decent place to start. But the real insights come from exploring the edges in completely different disciplines. How are hospitals thinking about this? How are logistics companies? What's working in education? The patterns that show up in vastly different organizations are the ones worth paying attention to.

Christensen's work explains why competitor-watching fails. Incumbents focus on sustaining innovations, the improvements their best customers demand. They miss the moves happening below them because those moves initially look too small, too cheap, too unsophisticated to matter. By the time they're big enough to notice, the game has already shifted.

The same thing is happening inside companies right now. The employee who built a workflow with ChatGPT to handle routine customer inquiries isn't on anybody's strategic plan. They're not presenting at the leadership offsite. They're in a cubicle, solving a problem with a free tool because the system they were given doesn't work. And that solution, messy and unsanctioned as it is, contains more intelligence about where AI actually adds value than most consulting reports.

What the Research Says About Bottom-Up Innovation

MIT's "GenAI Divide" report documented something that confirms everything I see in the field. The 5% of companies that succeeded with AI shared a common trait: they empowered line managers to drive adoption, not central AI labs. Tools purchased from specialized vendors succeeded about 67% of the time. Internal builds succeeded only a third as often.

The difference: vendor tools were shaped by real user needs. Internal builds were shaped by IT specifications. The users knew what worked. The architects didn't.

An academic research paper published in 2025 found that employees use generative AI roughly 3x more often than their managers estimate. Early studies linked unsanctioned ChatGPT adoption to higher individual creativity. The researchers described it as "shadow user innovation" and argued that firms should treat it as decentralized R&D capacity rather than a compliance violation.

KPMG's 2025 report framed it this way: shadow AI isn't a fringe issue. It's a signal that employees are moving faster than the systems designed to support them. The recommendation wasn't to crack down. It was to create safe environments for fast, creative experimentation.

Fujitsu put it even more bluntly: "The people experimenting with unapproved AI are the same people who will accelerate your AI programs tomorrow. Trying to ban shadow AI is like trying to ban Googling in 2002."

What To Do Instead:

The question for leaders isn't "how do we disrupt our industry?" It's "do we know what our people are already building? And have we created the conditions for those experiments to surface?"

1. Find your canaries. Identify the people in your organization who are already experimenting with AI. The ones building workarounds. The ones whose output quality jumped and nobody knows why. These people have already done the discovery phase for you. They know where the friction is because they live in it. Ask them what they've built. Ask them what works. Ask them what they need.

2. Run a shadow AI audit. Not to punish people, but to understand what's happening. What tools are in use? What problems are they solving? What data is being shared? BlackFog's data shows 43% of companies have no policy on AI tool usage at all. Start with visibility. You can't manage what you can't see.

3. Create a safe space, not a committee. The instinct is to form an AI steering committee. Don't. By the time the committee meets 3 times, your canaries will have solved 10 more problems. Instead, create a sandbox. Give people a secure environment to experiment with approved tools and real business problems. Set guardrails around data. Let them run.

4. Watch for patterns, not tools. The specific AI tool your employees are using doesn't matter much. It'll be different in 6 months. What matters is the pattern: which problems are people trying to solve? Where is the friction so bad that workers are willing to risk their security to fix it? Those patterns tell you where the real opportunities are. That's your roadmap.

5. Amplify what works. When a canary finds something that produces results, make it visible. Share it across teams. Scale it with proper governance. This is how bottom-up innovation becomes organizational capability. Not through mandates, but through demonstrated wins that others want to replicate.

6. Explore adjacent industries. Once you've mined the signals inside your own building, look sideways. Not at your direct competitors. At completely different industries solving similar structural problems. How logistics companies handle routing and scheduling might teach your service business something about workflow optimization. How hospitals manage triage might teach you something about customer prioritization. The patterns transfer. The tools are secondary.

The Conversation You're Not Having

In every organization I've worked with, there are people who've already started building the future. They're not waiting for permission. They're not waiting for a strategy. They're not waiting for a steering committee to approve a vendor.

They're solving problems. Every day. With tools they found on their own.

The question isn't whether to disrupt your industry. The question is whether you're paying attention to the disruption already happening inside your walls. The canaries are singing. The snow is melting from the edges.

Are you listening?