Schumpeter, AI, and the Art of Creative Destruction
Why artificial intelligence is not just coming for jobs, but for business models, value chains, and product assumptions.
Schumpeter, AI, and the Art of Creative Destruction
Why Artificial Intelligence Is Not Coming for Your Job. It Is Coming for Your Business Model.
Every generation believes it is living through unprecedented technological change, and most generations are at least partly wrong. The Victorians had railways, the early twentieth century had electricity, the post-war era had television, and the late twentieth century had personal computing and the internet. Today, we have artificial intelligence.
The technology is different, but the pattern is familiar. One economist understood this pattern better than almost anyone else: Joseph Schumpeter. While many economists focused on equilibrium, stability and efficiency, Schumpeter focused on disruption. He believed capitalism did not advance smoothly, but through upheaval, waves of innovation and the destruction of older ways of creating value.
His phrase for this process remains one of the most powerful ideas in economics:
Creative destruction.
If there is one concept that helps explain AI, it is this one.
The World’s Most Misunderstood Economic Idea
The phrase creative destruction sounds dramatic because it is dramatic. Schumpeter argued that innovation does not simply create new things; it also destroys old things. The motor car created enormous value, but it also destroyed large parts of the horse economy. The internet created trillions of pounds of value, but it also destroyed newspapers, travel agencies, classified advertising and countless other business models.
Innovation is rarely purely additive. It is often replacement. The mistake many organisations make is assuming AI will simply make existing businesses more efficient, when history suggests something more disruptive. AI is unlikely to merely improve industries; it is more likely to reshape them.
The Lessons of the Horse
A famous observation about technological change concerns horses. In 1900, horses were everywhere because transport, agriculture and commerce all depended on them. If you had asked an intelligent observer how to improve transport, they might have suggested better breeding, better feed, better roads or better stables.
What they probably would not have predicted was the disappearance of the horse economy itself. The automobile did not make horses obsolete overnight, but over time it changed the economics of transport so dramatically that entire industries disappeared.
This is Schumpeter’s point. The greatest threat often comes not from existing competitors, but from entirely new ways of solving problems. AI may represent a similar shift.
Why AI Is Different
Many previous technologies automated physical labour, whereas AI automates aspects of cognitive labour. That distinction matters because organisations have historically treated expertise as scarce. Knowledge workers existed because knowledge itself was expensive, information was difficult to access, analysis took time and decision-making required specialists.
AI alters those economics. Knowledge is becoming cheaper, analysis is becoming cheaper, content creation is becoming cheaper and software development is becoming cheaper. This does not mean expertise disappears, but it does mean its value may shift.
Whenever value shifts, Schumpeter’s process begins.
The Death of SaaS?
Every week, somebody confidently announces that AI will kill SaaS. Perhaps it will, perhaps it will not, but history suggests the answer is likely to be more nuanced.
Electricity did not kill manufacturing, the internet did not kill commerce and cloud computing did not kill software. They changed where value was created. AI is likely to do the same.
The question is not simply whether AI will replace software. The better question is which parts of software become commodities. If intelligence becomes abundant, then some traditional advantages may weaken. User interfaces may matter less, data entry may matter less and some types of workflow design may matter less.
At the same time, other capabilities may become more valuable: trust, context, governance, integration, proprietary data, industry expertise, accountability and distribution. Value migrates. It rarely disappears.
The Innovator’s Trap
One of Schumpeter’s most important observations is that successful organisations often struggle to adapt, which seems counterintuitive because successful companies usually have more resources, more customers and more talent. Yet history repeatedly shows that incumbents often struggle during periods of disruption.
The reason is that successful organisations are optimised for the existing model. Their processes, incentives, culture and metrics all reflect yesterday’s success. AI creates precisely this challenge.
The danger is not necessarily technological inability. The danger is organisational inertia. Companies continue optimising a model that is gradually becoming less relevant. They improve the horse while somebody else builds the car.
Product Managers and Creative Destruction
This creates an uncomfortable challenge for product managers because most product management focuses on optimisation. Better workflows, better usability, better reporting and better integrations are all worthwhile activities, but Schumpeter reminds us that optimisation is not always enough.
Sometimes the product itself is being challenged. Sometimes the market is changing. Sometimes the assumptions underneath the roadmap are becoming obsolete. The best product managers therefore operate on two levels.
The Operational Level
At the operational level, the product manager asks how the product can be improved. This protects today’s revenue by focusing on delivery, usability, performance, customer satisfaction and incremental advantage.
The Strategic Level
At the strategic level, the product manager asks what happens if the entire category changes. This protects tomorrow’s relevance by forcing the organisation to examine whether the assumptions underneath the product still hold.
Both levels matter, but only one is usually visible in the roadmap.
AI Specification Assistants: A Personal Example
Consider the growing use of AI in requirements engineering. Most organisations initially ask how AI can help them write specifications faster, and that is a sensible question. Schumpeter, however, would encourage a different perspective.
What if specifications themselves are an artefact of a previous economic reality? Historically, specifications existed because knowledge transfer was expensive. Business users understood the problem, analysts documented it, developers interpreted it and testers verified it. AI potentially changes that equation.
The opportunity may not simply be faster specifications. It may be an entirely different model of organisational knowledge. That is the difference between optimisation and creative destruction.
What This Means for Product Managers
Schumpeter’s work creates a difficult but useful discipline for product strategy. Before asking how to improve a product, product managers should ask which parts of the product category are becoming less valuable. Before asking what features should be added, they should ask what customer problem may be solved in an entirely different way. Before asking how to defend the current position, they should ask what assumption would make that position fragile.
This is where creative destruction becomes practical. It is not just an economic theory; it is a strategic warning. Product managers need to watch for moments where customer behaviour changes, costs collapse, scarce capabilities become abundant, new entrants solve the same problem differently, existing workflows become unnecessary, or value migrates from product functionality to data, trust, governance or integration.
Those are the moments where categories start to shift.
The Human Question
Whenever AI is discussed, conversations quickly drift toward replacement. Which jobs disappear, which roles survive and which industries change are all important questions, but they are not the only questions.
Schumpeter’s work reminds us that economic change creates winners and losers simultaneously. New opportunities emerge, new industries form and new capabilities become valuable. The challenge is not preventing change; it is adapting to it.
As Charles Darwin supposedly observed:
It is not the strongest that survive, nor the most intelligent, but those most responsive to change.
Whether Darwin actually said it is debated. The principle remains sound.
Final Thoughts
Joseph Schumpeter understood something that many organisations still struggle to accept. Innovation is not primarily about improvement; it is about replacement. New technologies rarely arrive politely. They reshape industries, alter incentives, move value and destroy some things while creating others.
Artificial intelligence appears to be following the same pattern. The companies that succeed will not necessarily be those with the biggest AI budgets. They will be those willing to question their assumptions, rethink how value is created and recognise when optimisation is no longer enough.
In other words, they will understand Schumpeter’s central lesson. The future is not built through continuous improvement alone. Sometimes it arrives through creative destruction, and history suggests it rarely asks for permission first.