Six months that ended a business
I worked at an online education company that served millions of college students worldwide. We had built a solid marketplace connecting students with a large library of content: material from professors, peer notes, research papers, and tutoring support. Our legal department worked hard to stay compliant across different countries and institutions. The business was growing steadily and predictably.
Then AI arrived.
Students no longer needed to research multiple sources. They had one chat window that could synthesize information instantly. They didn't need tutors for most questions, and they didn't even need to write original content anymore. A technology that barely existed six months earlier was turning our whole value proposition into a commodity.
To the company's credit, we did experiment with new business models. We built our own AI layer that could consolidate and personalize our content for specific student needs. We already had data scientists, machine learning experts, and AI developers who were excited to build for the new situation.
But the executive team wasn't prepared for disruption this deep. They had built their careers on decades of steady, predictable growth, and an existential threat was beyond what they knew how to adapt to. Leadership made one big bet. We got a chance to iterate and show some progress, but it wasn't fast or deep enough to compete with the worldwide forces against us. When the bet failed to turn the business around, leadership was out of ideas and patience. They shut the business down.
For those of us who wanted to compete, it was a devastating missed opportunity. I've thought a lot about what we could have done differently, and almost none of it could have started after AI arrived. By then the clock was already ticking. The work needed to happen earlier, while things were still good.
Scenario planning, done a year early
Step 1: Watch the signals
Scenario planning starts with a structured scan across Political, Economic, Social, Technological, Legal, and Environmental factors, usually called a PESTLE analysis. We were informally tracking competitors, but nobody was systematically watching technology for threats to our core model. AI-generated content was already showing up in adjacent spaces. Students were already changing how they researched and studied. Those signals were there if we'd been looking for them, and a structured scan would have flagged AI as a critical trend for our marketplace well before ChatGPT made it obvious to everyone.
Step 2: Three or four futures, including the ugly ones
With those signals in hand, we could have built 3-4 distinct scenarios. What if AI adoption accelerates and students stop needing curated content libraries at all? What if AI becomes a supplement that makes our content more valuable? What if regulators crack down on AI in education, and trusted, human-curated content becomes a premium product for a while? Each of those was plausible. Our leadership only knew how to plan for one future, continued steady growth, and when that future disappeared they had nothing to fall back on.
Step 3: Decide the triggers in advance
The payoff comes from writing strategies ahead of time and agreeing on the triggers that set them in motion. We could have picked core bets that held up in every scenario, like investing in our AI talent and building personalization. For the more disruptive scenarios, we could have had contingency plans ready with clear signals: if student usage of AI tools hits X%, launch the AI-powered content product. If content upload volume drops by Y%, speed up the pivot to personalized learning. What we actually did was make one big bet after the crisis hit, and when it didn't work fast enough, leadership had nothing left to try.
Agreed triggers also change who moves the roadmap. In Product-Led Alliance's State of Product Management Report 2026, 60.2% of product professionals said leadership escalations or new directives were the leading reason their priorities changed. A trigger you agreed on a year ago gives you a reason to change course that comes from the market, and it gives leadership something to check before they escalate.
We had everything but the practice
The infrastructure, the talent, the brand, and the customer relationships were all there. So were the data scientists and AI developers who were eager to build. What we lacked was a leadership team that had already thought through disruption before it arrived.
I think of scenario planning as rehearsal more than forecasting. You run it while things are calm, so that when one of the bad futures shows up, your team has already argued about it once and knows which plan comes off the shelf.