One of the most dangerous sentences in business is, “That’s never happened before.”
For most of my career, precedent has been one of the most useful tools in the room. You look at what happened last time, study what worked, understand what failed, and use that experience to make the next call. Good leaders do this constantly.
But precedent only works when the conditions are comparable. Right now, a lot of them aren’t.
AI is changing the economics of how companies compete, how quickly markets move, and how much can happen between one planning cycle and the next. Experience still matters enormously. But leaders need to know when they're learning from the past and when they're letting it limit what they can see.
There might be a very good reason nobody's done it before. But there's also a chance nobody's figured out how.
When Experience Becomes a Blind Spot
Businesses are built to learn from what already happened. We forecast revenue from previous quarters, allocate capital toward proven opportunities, and build strategies around patterns we’ve seen before. You don’t get smarter by ignoring the lessons other people have already learned.
Precedent becomes limiting when we start treating it as a requirement. If something has to resemble a previous success before we’re willing to take it seriously, we’re going to miss some pretty big opportunities.
How AI Is Changing Business Assumptions
AI is changing how much work a small team can take on and what it costs to test an idea. These shifts can alter which businesses are viable and how quickly new competitors emerge.
It’s getting harder to forecast the next five years by studying the last five. A business model that didn’t make financial sense five years ago might be viable today. But I want to understand what has actually changed. Lower delivery costs might make a business more viable, but they won't solve a lack of customer demand.
Which Lessons Still Apply
We need to look beyond what happened last time and understand why it happened, including whether the same conditions are present today.
The longer you operate, the more patterns you recognize. You learn what healthy growth looks like, which numbers deserve attention, and when a seemingly great result might be hiding a problem somewhere else. You get better at recognizing the underlying dynamics of a business, even when the circumstances are unfamiliar.
AI can help us examine more information and test those assumptions faster. But somebody still has to decide whether the comparison makes sense and whether the evidence supports the decision we're about to make.
Familiar patterns can be misleading, especially when they give us more confidence than the evidence warrants. We need to pay just as much attention to what's different as we do to what's familiar.
Being New Doesn’t Make Something Right
Of course, a changing market can also give us an excuse to believe what we want to believe. I’ve made bets that looked better going in than they did coming out. Novelty isn’t evidence, and being early doesn’t mean you were right.
Take Klarna. In 2024, the company announced that its AI assistant was doing work equivalent to 700 full-time customer service agents. That gave other companies a reason to reconsider their own staffing assumptions.
But in 2025, Klarna CEO Sebastian Siemiatkowski acknowledged that the company's emphasis on cutting costs had gone too far. Klarna began moving toward a model that offered customers more access to human support.
At the time of that shift, Klarna said its AI still handled two-thirds of customer inquiries. The efficiency gains remained substantial, but they hadn’t settled the question of service quality. Customers still needed access to someone who could help when the automated response wasn’t enough.
The lesson for me is to keep testing the decision after making it. A new capability can change the economics of a business while leaving familiar obligations to customers firmly in place. If the evidence shows we’ve put too much weight on cost and too little on quality, we need to adjust.
Evaluating Businesses Without a Clean Comparison
When we’re evaluating a business at RAD Intel, I want to understand what has limited its growth and whether we have a credible way to address it.
If serving more customers has always required hiring more people, could technology change how some of that work gets done? If the company has deep expertise but limited distribution, could relationships elsewhere in our portfolio help it reach new customers?
Those are possibilities we need to examine carefully, including the investment required, what would need to improve, and what we could test before committing to a larger plan.
Historical performance helps us understand the business we’re buying. We also need evidence for what we believe we can help it become. The more ambitious that belief, the more specific I want us to be about how we’ll test it.
Before It Becomes Obvious
There's comfort in waiting for precedent. Once the market is established, the technology has matured and enough companies have proved the model, the decision becomes easier to defend. It also becomes a different decision.
The past should make us smarter about the future, not limit what we believe is possible. We're going to operate through a period where some of the assumptions leaders have relied on for years stop holding. The companies that navigate it well will be the ones that know which lessons still apply, which assumptions need updating, and when something genuinely new is happening.
So when someone says, “That's never happened before,” I'm going to ask why. I've heard plenty of good reasons to walk away from an opportunity. A lack of precedent, on its own, isn't one of them.
What Marketers Should Know Now
Investors are betting on AI to transform established businesses. A new investment firm called Aldaron has raised approximately $885 million for its debut fund, with plans to acquire companies and improve their operations using AI. Its approach includes embedding engineers directly into portfolio businesses to drive growth. It's an interesting example of investors looking beyond historical performance to what a company could become. Read The Wall Street Journal's coverage
Companies are seeing AI returns but struggling to scale. An October study from BearingPoint found that nearly three-quarters of surveyed companies reported positive financial results from AI, yet fewer than a third had moved beyond pilot projects. Regulatory requirements and legacy technology remain major obstacles. The findings highlight the gap between proving a technology works and building a business around it. Read Reuters' coverage
AI is changing what it takes to build a company. Startups are using AI to operate with smaller teams, challenging traditional assumptions about the staffing required to grow. Some are generating substantial revenue with relatively few employees, although questions remain about how well those models will hold up as companies scale. Read The Wall Street Journal's reporting
WPP is rethinking the traditional agency partnership. On October 6, WPP and technology services company Kyndryl announced an expanded partnership that includes AI-driven operational transformation, joint business development, and a broader marketing relationship. The agreement reflects how agency relationships are evolving as technology becomes more deeply integrated into service delivery and growth strategies. Read the announcement
