The Innovation Imperative, Revisited: Why “Disruption” Still Gets Misunderstood
Over six years ago, I wrote about disruptive innovation. The piece was about why most people (CXOs, consultants, shop-floor managers alike) throw the term around without really understanding it. Reading it back today, in a year when every second LinkedIn post claims AI will “disrupt everything,” the core lesson feels more relevant than ever. Innovation and disruption are a process, not a product.
And I felt that may be refresher update will be good as I still see people chasing shiny toys as world changing innovation.
Go back to the classics. Google didn’t start as a search company. It started as two PhD students trying to map how web pages link to one another. The search engine was a by-product of that experiment. Airbnb began with an air mattress and a free breakfast, meant to help two guys make rent. Uber launched with three cars on an app. None of these set out to build a “disruptive product.” They stumbled into a business model that happened to scale.
That distinction, process and business model over product, is exactly what’s being missed in the current AI gold rush. Plenty of companies are racing to bolt a chatbot onto their product and calling it disruption. It isn’t. It’s a feature. Real disruption shows up when the business model changes: how value is created, delivered, and paid for.
We’re actually watching this play out in real time. Generative AI reached roughly 2% share of the enterprise software market in about a year. McKinsey estimates that’s three times faster than it took SaaS to hit the same milestone a decade earlier. That’s not because the AI is a shinier product than SaaS was. It’s because the underlying economics of building and delivering software are being rewritten: smaller teams, faster iteration, usage-based pricing replacing per-seat licensing. That’s business-model disruption, not feature disruption.
I used this example before and it still holds. In 1886, Carl Benz patented the first practical automobile. It was genuinely brilliant engineering. And yet it changed almost nothing, because early cars were luxury toys for the wealthy. Disruption of the transportation market didn’t happen until Ford’s moving assembly line made the Model T affordable enough for ordinary households. Between 1908 and 1927, Ford built more than 15 million of them. The invention wasn’t the disruptor. The manufacturing and pricing model was.
Every technology wave replays this pattern. The breakthrough capability arrives first, and it’s genuinely impressive. But it stays a curiosity until someone works out how to make it cheap, accessible, and embedded into daily workflows. That’s exactly the phase a lot of enterprise AI is in right now: the “Benz patent” stage, not yet the “assembly line” stage. The companies that figure out the packaging, pricing, and distribution (not just the model quality) will be the ones that actually disrupt their markets.
Kodak invented digital photography and still went bankrupt in January 2012. It was undone by a business it was too invested in the old model to cannibalize. That story is now well over a decade old. But ask any large enterprise today how they’re approaching AI, and you’ll hear an echo of the same pattern. There’s plenty of investment in “incremental” innovation, a copilot here, an automation there, because it’s easy to fund and easy to measure. It doesn’t require anyone to admit the core business model might need to change.
Get Srinivas Kollur’s stories in your inbox
Join Medium for free to get updates from this writer.
Incremental innovation isn’t wrong. It’s necessary, and most large organizations are genuinely good at it: new features, service improvements, staying visibly current. But it’s not the same muscle as disruptive innovation. Disruptive innovation requires tolerating ambiguity, ceding control to small autonomous teams, and sometimes cannibalizing a profitable product line before a competitor does it for you. Large companies with the most resources are often the ones most institutionally allergic to doing this. Not because they lack talent, but because they have too much to protect.
By 2026, Gartner expects the large majority of enterprises to have generative AI in production, up from a small single-digit percentage just a few years earlier. Adoption is no longer the differentiator. What separates the disruptors from the disrupted now is the same thing it always was: who’s willing to change the business model, not just the toolset.
There’s one filter I still find useful. Don’t start with “what can this technology do?” Start with “what experience (for a customer or an employee) is genuinely broken or frustrating right now?” Then work backward to whether an incremental fix or a fundamental rethink of the business model is what’s actually needed. Uber, notably, wasn’t a textbook disruptor by this definition either. It didn’t create a new market or a radically lower price point. It just made an existing experience (hailing a cab) dramatically better. That’s a legitimate and valuable form of innovation too. Blinkit, Instamart and several others quick commerce players also innovated on the buying convenience beyond what was actually presented as a tactical social opportunity presented during Covid. These aren’t the same species as what happened to Kodak or Blockbuster.
The lesson from 2020 hasn’t changed. It’s just more urgent now. Innovation earns its “disruptive” label only when it changes how value gets created and captured, not when it changes what’s on the feature list. As every industry rushes to bolt AI onto existing products, that distinction is worth holding onto.
Sources
- Kodak Chapter 11 bankruptcy filing, January 19, 2012. U.S. SEC Form 8-K: https://www.sec.gov/Archives/edgar/data/31235/000119312512016443/d285282d8k.htm
- Ford Model T production history (15M+ units, 1908 to 1927). Ford Motor Company corporate history: https://corporate.ford.com/articles/history/the-model-t/
- Generative AI enterprise adoption pace vs. SaaS. McKinsey & Company estimate, cited in SaaS Mag: https://www.saasmag.com/how-gen-ai-disrupts-saas/
- Gartner forecast on enterprise generative AI production use by 2026: https://www.scrumlaunch.com/blog/ai-in-business-2026-trends-use-cases-and-real-world-implementation
- Original 2020 article for reference and context: https://www.wipro.com/infrastructure/the-innovation-imperative-how-to-stay-ahead/
The gap between what IT Services companies design and what customers actually need is where value gets left on the table. I help IT services and infrastructure leaders close that gap by putting customer reality at the center of strategy, not at the end of it.
