Creative Destruction in the Age of GenAI

A Nobel Lesson for the Legal Industry

When Joel Mokyr, Philippe Aghion, and Peter Howitt received the 2025 Nobel Prize in Economics, headlines focused on their ideas about growth and technological progress. But behind the celebration lies a deeper truth: real progress comes through creative destruction - the renewal that happens after something better has come about.

That insight hits close to home for law firms exploring Generative AI. Yes, adoption numbers look promising: 79% of legal professionals report some AI use. Yet only 31% use it personally, and fewer than one in five firms have rolled it out firm-wide. Even more telling, nearly 60% of those who have adopted it don’t measure return on investment.

In other words, the legal sector isn’t transforming - it’s experimenting without accountability.

Why Real Innovation Feels So Hard

Buying a new tool is easy. Changing how work actually gets done is not.

Many firms are trying to plug AI into systems built for a different era - the billable hour, risk-averse policies, and deeply ingrained workflows. It’s innovation by addition: layering something shiny on top of something rigid and calling it progress.

But creative destruction doesn’t start with tearing things down or by incrementally improving what already exists. It starts when something demonstrably better takes root. Startups understand this instinctively. When digital cameras outperformed film, the darkroom vanished. When Westlaw and LexisNexis outshone law libraries, shelves of casebooks gave way to screens.

Inside established firms, though, the order often flips - teams incrementally improve and/or dismantle what exists before the replacement is ready to stand on its own. The latter is not creative destruction - it’s creative chaos.

How to Build the New Beside the Old (Without Breaking What Works)

Creative destruction isn’t a natural evolution - it’s a managed decision. It hurts. Something you’ve invested in, that people like and that’s still “working,” has to make room for something better. That’s not failure; it’s leadership.

Aghion, Howitt, and Mokyr show that growth happens when the new can replace the old once it proves more efficient. But replacement never happens by itself - it requires design: incentives that reward learning, protection for early experiments, and the courage to stop doing what no longer serves the goal. As Joel Mokyr notes, progress flourishes where experimentation is encouraged and failure isn’t punished- but even then, someone has to choose.

That’s the core organizational challenge for law firms. The issue isn’t the tech - it’s the architecture of how decisions, risk, and work flow. You need a structure that lets you explore tomorrow while still delivering today. Researchers Michael Tushman and Charles O’Reilly call this organizational ambidexterity - the ability to exploit what works while exploring what’s next. It’s not a personality trait; it’s a system you build.

A candid example from O4I.
We faced this ourselves. Our Steering Wheel platform was delivering value and had traction. Yet we realized it was getting in the way of what clients needed most right now: disciplined GenAI adoption that actually changes behavior and produces ROI. We tried to do both - and it split focus, slowed learning, and dulled impact. So we made the hard call to sunset the platform and commit 100% to GenAI adoption. It was painful. It also unlocked speed, clarity, and better outcomes for clients.

That’s what creative destruction looks like up close: not a clean swap, but a deliberate trade-off.

For firms, the same logic applies:

  • Design for coexistence, not comfort. Let the new grow beside the old with explicit milestones and a clear “earn the right to replace” path.
  • Protect the experiment and the core. Keep client delivery stable while carving out time, talent, and metrics for the new.

Decide on purpose. Sunsetting the old is a decision made with evidence and proof of concept, not drift.

Three Ways to Turn GenAI Adoption Into Measurable ROI

  1. Start With Discovery, Not Deployment
    Before investing in another tool, talk to clients. What do they actually value -  speed, cost control, or confidence in accuracy? Use structured discovery interviews to surface what matters. When the job-to-be-done is clear, GenAI’s role becomes obvious.
  2. Run Small, Accountable Experiments
    Pick one use case - a research memo, a contract review, a client alert - and run a short trial. Define success before you start: time saved, quality improved, or satisfaction gained. Two-week cycles keep learning fast. A “failed” experiment that teaches you something valuable is still a return on investment.
  3. Plan for Discovery, Not Prediction
    Instead of long-range AI strategies, use discovery-driven planning. Write down assumptions, test the riskiest first, and adjust as you learn. It’s a disciplined way to reduce uncertainty while maintaining momentum.

The Timing Problem: When to Let Go

True creative destruction happens when the new becomes undeniably better than the old. Kill the old too soon, and you create confusion; protect it too long, and progress stalls. The best firms let the new grow beside the old until it earns the right to replace it - just as online research once coexisted with physical libraries before finally overtaking them.

That coexistence requires patience, data, and courage - the willingness to measure what’s working and to stop doing what isn’t.

Weekly GenAI Tip: How to Apply This Insight

This week, pick one process in your firm where AI is being tested but not tracked. Ask three simple questions:

  1. What problem was this tool supposed to solve?
  2. How do we know whether it’s working?
  3. What’s stopping us from measuring its impact?

You’ll often find that lack of ROI isn’t a technology problem - it’s a leadership one. Start by defining what success looks like, and you’ll create the conditions for creative destruction to happen - intentionally, not accidentally.

Stay Ahead of the Curve

If you found this perspective helpful, subscribe to Weekly GenAI Tips -  short, actionable insights on turning GenAI chaos into measurable ROI.

Get Weekly Gen AI Tips