Andy Mochan jumped.
Piper Alpha, July 1988. 167 men died on a North Sea oil rig that turned into a fireball in under two hours. Andy Mochan, a 35-year-old roustabout, ran to the edge of the rig and jumped more than 100 feet into freezing water because the alternative was burning alive. He survived.
The image of that jump traveled into every corporate strategy room in the 1990s because it was the perfect metaphor for change. Daryl Conner's Managing at the Speed of Change first introduced the metaphor. It was a single crisis. A single moment of decision. Jump or burn. We are no longer in the 1990s, though. We are in the AI era, where the fire moves faster than the smoke, and the rig you're standing on may already be structurally compromised in ways your dashboards aren't built to see. The burning platform is worth revisiting. Not to retire it. To rebuild it for the curve we're actually on.
The burning platform metaphor worked when corporate disruption was episodic. In the AI era, the window between weak signal and collapsed unit economics has closed to near zero. The new playbook is not to anticipate the fire. It's to start building the next platform now, while the cost of building it has collapsed faster than the cost of being late.
The fire moves at AI speed now.
John Kotter also took Piper Alpha and made it the foundational change management story for a generation of executives. It worked because corporate disruption in his era was episodic. A crisis emerged. A leader rallied the org. Transformation happened on a multi-year horizon, with multi-year warning. The whole frame depended on one assumption: that the gap between "platform starting to smolder" and "platform fully engulfed" was long enough for leaders to see it, communicate it, and mobilize around it.
AI breaks that assumption.
The window between "weak signal" and "your business model is gone" is collapsing toward zero in entire categories. Customer service teams have gone from 500 people to 50 in under 18 months at companies that didn't see the curve until it was past them. Legal research, design production, sales development, code generation, paralegal work, market research, junior analyst roles, early-stage venture diligence. Each one is a rig where the smoke and the fire arrived in the same quarter. The leaders who relied on the old anticipation logic kept scanning the horizon for one burning platform and missed that the entire field of platforms had shifted underneath them.
This is why the metaphor needs an update. Not because it was wrong. Because the time constant changed. The same instruction that told a CEO in 1995 to start a transformation program once the smoke was visible now tells a CEO in 2026 to launch a five-year strategic refresh while their unit economics are being rewritten in real time. Same instruction, completely different physics.
You can't see this fire coming with the old instruments.
The old playbook said: scan for emerging burning platforms, act early. That assumed signals would arrive on a roughly linear curve and that internal teams could mobilize fast enough to respond once the signals were clear. Neither of those is true anymore.
AI disruption is non-linear. It compounds in capability jumps that don't show up as quarterly headwinds until they show up as collapsed unit economics. A 30% productivity improvement on a key workflow doesn't matter much until it becomes a 10x improvement, and by then it doesn't matter because your competitor's pricing already moved. The instruments most corporates use to watch for threats were built to track linear curves. They are not the instruments you need for this fire.
The deeper problem is structural. Even when corporate leaders see the curve, the internal teams that would have to build a response are operating on incentives, timelines, and tooling that were designed for the pre-AI rate of change. Asking a quarterly-managed P&L team to build a venture that uses AI to reach unit economics no incumbent can match is not a culture problem. It is a structural mismatch. The instruments are wrong, the team is wrong, the timeline is wrong, the incentives are wrong. You don't fix that by telling people to move faster.
Build the parallel platform. Build it with AI.
The structural answer to slow-burn disruption has always been the same: build the next platform in parallel to the current one, on different incentives, outside the walls of the current business. Strategy researchers have written about this kind of build as dual transformation. The core takes care of the present. The other side builds what's next. That's the core logic of corporate venture building. You don't fix the rig that's smoldering by holding more meetings on the rig. You start the next rig somewhere else, on different ground, while you still have the cash flow and brand and customer access from the original one.
What's new in 2026 is that the act of building the parallel platform is itself now accelerated by AI in ways that change what's possible.
We wrote about this in our essay on our Agentic Venture Studio, where our team is building nine companies simultaneously with thousands of AI agents. Agentic systems are doing the work of research, design, build, and go-to-market in parallel. We are running nine concept-to-pilot tracks simultaneously inside that model. The cost curve of building the next platform has collapsed faster than the cost curve of being late to it.
OneHealth Studio is a working example of what this enables. An AI-native operating model from day one. Not a legacy stack retrofitted with a copilot. Not a digital program bolted onto a 30-year-old core. A company that uses AI as the substrate of how the business actually runs, built outside the walls of the corporate parent that needed the new platform but couldn't have built it from inside.
The implication for corporate leaders is bigger than tooling. It means the parallel platform you needed two years to build can now be built in a quarter or two. The capital required is smaller. The team required is smaller. The internal coordination overhead is smaller. That changes the math on when you should start. The right move is no longer to wait for the smoke. The right move is to start building the next platform while this one is still profitable, because the cost of building it has collapsed faster than the cost of being late.
Your real smoke signals.
The old diagnostic was: where is the fire likely to start? The new diagnostic is sharper, and it's the one I'd ask if I were sitting across from you at your strategy offsite.
Which of your revenue lines has the highest exposure to AI-driven unit economics collapse, and how fast? Not "could AI affect this," but "what does this line look like if a competitor reaches our gross margin with a fraction of our headcount." Which capability would you green-light an outside team to build with agentic systems tomorrow if you didn't have to defend last quarter's number? Where is a smaller competitor making a move with AI right now that doesn't make commercial sense yet but will in 18 months?
If those questions land and you don't have answers, the platform is already smoldering. The burning platform was always a lagging indicator. In the AI era, it's a lagging indicator on a much shorter clock. By the time you smell the smoke, you needed to have jumped a year ago.
Better move. Start the next rig now.
.avif)
