Six months ago, my team and I embarked on a project. We were tired of the endless stream of articles about AI that were long on hype and short on data. We wanted to know what really makes an AI team successful. Not the fluff, but the hard, data-backed truth.
So we analyzed 500 AI teams. Some were at Fortune 500 companies, others were scrappy startups. We looked at their funding, their tech stack, the credentials of their team members, and of course, their project outcomes. We crunched the numbers, ran regressions, and spent hundreds of hours interviewing team leads and executives.
The results were not what we expected.
Everyone thinks success in AI is about having the most PhDs from Stanford, the biggest GPU cluster, or the most sophisticated models. Those things help, don’t get me wrong. But they are not the primary predictor of success. Not even close.
The Usual Suspects That Don't Matter as Much as You Think
When we started, I had my own hypotheses. I’ve been building and investing in companies in Silicon Valley for over a decade. I’ve seen teams with every advantage fail, and teams with nothing but a crazy idea succeed. I thought maybe it was about raw talent density. Or the budget. Or access to proprietary data.
But the data told a different story. We saw teams with nine-figure budgets get lapped by startups with a seed round. We saw teams full of Kaggle grandmasters produce nothing of value, while teams of relatively unknown engineers shipped game-changing products.
It turns out that while resources and talent are correlated with success, they aren't the cause. They are often just a proxy for something much more fundamental.
The Surprising Truth: It’s About the Leader
The single most important factor that predicts whether an AI team will succeed or fail is the leader's ability to make decisions under extreme uncertainty.
That’s it. It’s not about their technical chops, their MBA, or their network. It’s about their gut. It’s about their willingness to make a call when they have 40% of the data they wish they had. It’s about their courage to be wrong.
AI is not like traditional software development. The roadmap is not clear. You are exploring, not executing a known plan. You are dealing with probabilities, not certainties. In this environment, a leader who waits for all the information will wait forever. The team will get stuck in analysis paralysis, endlessly tweaking models and running experiments, never shipping anything.
I remember one of my portfolio companies, a small AI startup trying to disrupt the logistics industry. They had a brilliant team, but they were spinning their wheels for months. They had two potential paths for their core algorithm. Path A was more academically interesting but technically challenging. Path B was less elegant, but simpler to implement. The team was split. The CEO, a first-time founder, was afraid of making the wrong choice. He wanted more data, more proof.
Finally, after a tense board meeting, I told him, "Pick one. I don't care which one. Just pick one and go. If it's the wrong choice, we'll know in three months and we'll pivot. If you don't pick, we'll be dead in six."
He chose Path B. It wasn’t perfect, but it was good enough. They shipped the product. They got their first customers. They learned from real-world feedback. That company was eventually acquired. The CEO later told me that that was the moment he truly understood his job wasn’t to be right; it was to make decisions so the team could move forward.
Why This Kind of Leadership is So Rare
In most corporate cultures, being wrong is punished. You get dinged on your performance review. Your project gets cancelled. You lose political capital. As a result, managers are incentivized to avoid risk. They create endless committees, demand exhaustive reports, and generally cover their asses.
This is death for an AI team.
An AI team needs a leader who provides air cover. A leader who says, "We are going to take this hill. The path is foggy, and we might have to retreat and try another way. But we are going to try. And if we fail, the blame is on me."
This creates psychological safety. It gives the team permission to experiment, to take risks, and to fail. And in AI, the only way to succeed is to fail, learn, and iterate, faster than anyone else.
How to Lead in the AI Era
So how do you become this kind of leader? It’s not something you learn in a book. It’s a muscle you build.
- Embrace the 70% rule. If you have 70% of the information, and you’re not making a decision, you’re being too slow. Make the call.
- Timebox your research. Give the team a fixed amount of time for exploration. It could be two weeks, it could be a month. But at the end of that period, a decision is made, and the team moves into execution.
- Celebrate learning, not just winning. When an experiment fails, what did you learn? Share it openly. The only true failure is the one you don't learn from.
- Be the "Chief Decider". Your team is looking to you to break ties and make the hard calls. That is your job. Don’t abdicate it.
I’ve invested in over 200 companies, including some of the biggest names in AI like Anthropic and OpenAI. I’ve seen this pattern play out again and again. The teams that win are not always the ones with the most resources. They are the ones with leaders who have the courage to navigate the fog of the unknown.
Stop looking for the perfect model or the perfect dataset. Look for the leader who is willing to be imperfect, to be decisive, and to lead their team into the future, one uncertain step at a time.
Frequently Asked Questions
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
Do all experts agree with this view?
No, and that's fine. The best ideas in business are often contrarian. I share my perspective based on my experience and data, but I encourage you to seek out opposing viewpoints and form your own conclusions.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.