The Cost of Building Collapsed. The Cost of Being Wrong Did Not.

Jun 18

Summary

AI is flooding organizations with new opportunities. Prototypes that once took months now take hours, but building something quickly was never the same as knowing it deserved to be built. The leaders pulling ahead with AI are rebuilding the decision discipline that the cost of execution once enforced automatically.

When Building Got Cheap, Deciding Got Hard

For decades, organizations were constrained by their ability to execute. Those constraints felt frustrating because they forced difficult conversations early in the process. Before leaders would commit significant capital, teams had to justify the spend.

This meant asking uncomfortable questions.

What problem are we solving?
Who are we solving it for?
What outcome are we trying to achieve?
Should we invest at all?

The cost of execution was never really the safeguard, it was a proxy for one. It forced the risk conversation by making you pay to get past it. Today, many of those constraints are gone, along with the conversations they once forced.

Moses Adedoyin, who leads venture and innovation at GuideWell, described what that shift looks like in practice. A clickable prototype that once cost roughly $100,000 and three months through an outside consultancy can now be produced in an afternoon for a few hundred dollars. That speed also removed the chokepoint that once forced your team to defend an idea before building it. When building took three months, you argued about the assumptions first. When it takes three hours, you just build, and discover what you assumed only after it's live. It feels like savings, but the speed only pays off if you learn something from it. Shipping an untested assumption at full scale isn't learning faster, it's just finding out later, with your brand on the line.

The bottleneck has moved from building to deciding, and the safeguards that once emerged naturally now have to be redesigned on purpose.

The New Leadership Challenge

In my conversations with leaders, I expected to hear countless complaints about AI models, agents, copilots, and integration challenges. Instead of talking about actual AI technology, they were all speaking about human decisions. The same ones now landing on your desk faster than you can weigh them.

Gina Mitchell, who led strategic engagements at Mastercard Digital Labs, spoke about how decisions at different ends of the spectrum can impact the day-to-day operations. “At one extreme, you can give your employees all the tools and tell them to go teach yourself, figure it out. At the other end, you can over-engineer the process of governance which just slows things down and discourages people from even attempting to experiment."

One end ships untested bets at speed straight to the customer and the other progresses so slowly that bets die on the vine. Neither is a strategy, and both are what you get in the absence of one.

AI Didn't Remove Uncertainty

AI can now generate ideas faster, create prototypes faster and automate workflows faster. What it can't do is decide for you whether an opportunity deserves investment. That is still a human judgment, but organizations no longer have to confront those decisions before they start building.

You can't improve the decision without first improving the case that lands on your desk. What struck me the most across these conversations is how similar the responses became once building got fast and cheap. The leaders creating value from AI don't start with technology, they slow down long enough to understand what needs to be true before additional resources are committed.

That sounds rather obvious in theory, but it isn't in practice.

"Let's define the assumptions at risk," is how Aman Shah, VP and Head of AI & Ventures at VNS Health, describes his ways of working. His model begins with the problem, not the technology. "I ask two questions all day," he told me. "What problem are you trying to solve? What outcome are you trying to achieve? Those two questions are actually hard." They’re hard because AI makes it easy to skip over them in the name of speed, without realizing how costly they can become unanswered.

They also spend time distinguishing between uncertainty that matters and uncertainty that doesn't. Not every assumption deserves the same attention. The leaders I spoke with were consistently focused on the questions that could make or break an initiative. Once building becomes inexpensive, understanding where a bet is most likely to fail becomes more important than understanding how quickly it can be built.

They create moments where evidence is reviewed before additional resources are committed. Moses described this pattern as “Any idea is already a solution... they're going to invest millions of dollars, and then they find out that it's not working." When building was expensive, budget cycles caught some of those bets early. Now nothing does, unless a leader builds in that check on purpose.

Cheap Building with AI Doesn't Lower the Stakes

The wrong bets still cost you, since your customers are exposed to them and trust is spent. What got cheap was the prototype and what stayed expensive is being wrong, with customers deciding on their own timeline regardless of how fast you deployed it. The strongest teams aren't waiting until after a large investment has been made to learn whether they were right, they're creating opportunities to learn.

The leaders pulling ahead with AI are creating new ways to force the conversations that used to happen naturally. They're not abandoning the fundamentals of decision-making, they're doubling down because AI made it so easy to abandon them.

The cost of building has collapsed, but the cost of building the wrong thing hasn't.

The leaders pulling ahead understand the difference.