Why All-Star Teams Fail in the AI Era: Executive Guide

I have learned that a core belief of mine is wrong: a room full of top performers doesn’t automatically create a team that thinks and performs well together.

I wish I could say this epiphany came to me naturally but it was actually the result of reading Jon Levy’s new book “Talent Intelligence” and watching the Netflix series “Boots”. Both address this idea but in different ways: one highly academic and the other through story telling.

If you have not had a chance to watch Boots, it tells the story of a platoon of recruits with no special talent. Each feels they are the best but have to learn to function as a unit by eliminating the weaker part of themselves. The unit survives a military training program that breaks individuals with superior raw ability. The difference isn't that they're smarter or stronger. It's that they learn to trust each other in ways they had not experienced before.

It was the combination of the book and the series that made me realize there is a dangerous assumption most companies still operate under: assemble the smartest people, and excellence multiplies.

In reality? Big egos collide. Power struggles emerge. Knowledge gets hoarded. Teams fragment. The all-star roster fails.

That's team intelligence. And it's about to become your competitive advantage or your biggest liability.

What does Netflix's Boots reveal about trust and team dynamics?

Netflix's military drama isn't subtle about this. The show alternates between drill instructors who use fear and hierarchy to control recruits, and instructors who build trust and allow teams to self-organize.

When authority relies on intimidation and withholding information, the platoon fragments. When it shifts to compassion and transparency, even for brief moments, the same group reorganizes almost instantly.

Cameron's journey as a closeted gay recruit navigating an institution built on conformity illustrates a deeper truth: psychological safety has the largest impact on minorities and marginalized team members. Harvard Business School research confirms this. diverse teams with psychological safety see enhanced collaboration across all dimensions.

By the final episodes, the Boots platoon accomplishes feats that individual superstars could not. One of the members of the troop struggling with an injury is supported by other team members who willingly take the additional load. The main character breaks rank to find another recruit who had wondered off and become injured.

The unit succeeded in their final test. Not because they became stronger or smarter. Because they learned to trust each other’s strengths.

Why Do High-Performer Cultures Systematically Fail?

Organizations have spent 30 years chasing "A players." Honestly, I was one of them and went so far as to write a book about it.

Venture capital firms stacked portfolios with top-tier talent. Tech companies fought wars over engineering superstars. Sports franchises mimicked the model: assemble Hall of Famers, win championships.

It didn't work. The reason: chemistry doesn't develop when high-performing egos compete for resources and recognition.

Don’t believe me? The 1992 Dream Team, arguably the greatest roster ever assembled, reportedly lost a practice game to college players.

Corporate America replicated the same model. High performers became untouchable superstars. Lower performers were managed out. The logic seemed sound: stack excellence, get excellence.

What researchers now reveal: excellence in isolation multiplies friction, not performance.

Who Are the Glue Players Your Organization Systematically Devalues?

Levy's research identifies a different archetype: the "glue player."

Glue players aren't the most visible team members. They don't generate the biggest deals, write the most impressive code, or close the highest-value clients. They make everyone else better.

They anticipate needs. They take initiative without seeking credit. They create the conditions for others to excel. They possess unusually high emotional intelligence. In corporate terms, they're the connectors, the mentors, the quiet leaders.

High-performer cultures systematically devalue them. They don't show up on individual performance metrics. They're invisible until they leave, at which point teams mysteriously lose cohesion despite having the same roster of "talent."

This is what Boots demonstrates visually. Cameron isn't the strongest recruit. Ray isn't the smartest. Ochoa isn't the fastest. But when they learn to support each other, and when Cameron shifts to enabling teammates instead of proving his worth, the platoon transforms.

That's team intelligence. Honestly, that’s also a better description of high performance.

How Does AI Actually Change the Performance Gap?

Here's the scenario that should concern every HR leader: in 18 months, the performance gap between your highest performer and your average performer will be narrower than ever before.

McKinsey research shows generative AI improves performance most dramatically for lower-skilled workers. (Brynjolfsson, Li, & Raymond, 2023) A study of GPT access for customer service agents found that the weakest performers showed productivity gains of 35-40%, while the best performers saw single-digit improvements.

This is capability democratization. It's the death of the high-performer culture as we know it.

Organizations that built competitive advantage through access to top talent could watch that advantage evaporate. Meanwhile, teams that cannot collaborate, harbor internal resentment, or operate in silos will find that AI magnifies dysfunction.

P&G's 2025 research is telling: teams working with AI were 12% faster.(P&G, 2025) But the combination of human collaboration plus AI created opportunities neither could achieve alone. Teams that already trust each other capture disproportionate AI gains. Teams that don't? They get efficiency improvements they can't capitalize on because they won't share knowledge or build on each other's ideas.

The gap isn't in AI tool availability. It's in social infrastructure.

What happened when ServiceNow abandoned formal AI training?

This is where theory meets operational reality.

ServiceNow, a 28,000-person enterprise software company, faced a familiar problem in November 2025: AI adoption was stalling despite massive investment in tools and training. Top-down formal training wasn't working. Neither were executive mandates. (Bloomberg, 2025)

So the company flipped the model entirely.

ServiceNow identified 1,000 employees who were already using AI effectively and possessed high emotional intelligence. They positioned them as "workplace influencers." The company gave them facilitation training and recognition, then stepped back.

The result: AI adoption accelerated faster than any formal training program could achieve. Peer-to-peer learning worked because psychological safety remained intact. Colleagues learning from colleagues at similar skill levels create zero status threat. No performance review implications. No shame in asking questions.

This is team intelligence operating at scale.

HubSpot, facing the same challenge, intentionally recruited both AI skeptics and AI champions as peer influencers. The insight was simple but powerful: trust is built when people see themselves reflected in the teachers.

Columbia University research supports this. Optimal peer learning happens in clusters of 4-6 people where psychological safety is highest and knowledge transfer is most practical. Small enough for psychological safety. Large enough to create diverse perspectives. (Columbia, 2025)

McKinsey's 2025 research on learning organizations is damning: Nearly half of employees want more formal AI training, but peer-learning adoption rates consistently outpace formal programs by 3-4 times. (McKinsey, 2025) In other words, organizations already have the solution. They just have to design for team intelligence instead of individual high performers.

Why Do High-Performer Cultures Block AI Adoption?

Here's what happens when you introduce AI into a high-performance culture.

Adoption becomes political. Who gets trained first? Whose workflow changes? Who gets credit for productivity gains? People hoard knowledge instead of sharing it. They position themselves above colleagues to maintain status. They're uncomfortable admitting gaps or asking peers for help because doing so signals weakness in a zero-sum game.

Meanwhile, organizations with higher psychological safety with cultures that value connectors and glue players experience employees who are eager to teach each other, collectively figure out how to integrate AI, and build on shared learning without fear of status loss.

The companies moving fastest on AI adoption aren't the ones with the best algorithms. They're the ones with the strongest peer networks and the most psychologically safe environments.

That's not a coincidence. That's the operational reality of team intelligence.

What Does This Mean for Your Organization Right Now?

The incompatibility is stark. High-performer cultures are built on individual achievement and competition. Team intelligence requires collaboration, psychological safety and more glue employees than high performers. You cannot optimize for both simultaneously.

If your team composition is all-stars with few connectors, your AI adoption will plateau. If your promotion system rewards individual metrics over collaboration, your peer networks will atrophy. If your culture punishes mistakes, people won't take the risks required to learn AI tools effectively.

ServiceNow's peer influencer model is an example of operational architecture. And companies that haven't built similar infrastructure will watch billions in AI investments deliver mediocre returns while competitors surge ahead.

Companies are already discovering that AI investments mean nothing if employees won't adopt the tools. Adoption happens only when teams have the social infrastructure to teach each other.

Organizations that spent decades building teams of high performers are sitting on a liability. They assembled rosters optimized for a world where individual excellence was scarce. Now, in an AI world where capability is abundant, those same teams are fragmented, political, and inflexible.

The solution is to focus on dismantling the myth that all-stars create all-star teams. It's building psychological safety and the ability to fail fast and learn faster. It's valuing glue players as much as superstars. It's also creating the conditions for collective intelligence instead of enabling individual heroics.

Most critically: the training crisis isn't just about AI literacy. The real problem is this is something most executives have never been trained to build, measure, or maintain.

How Do You Start Building Team Intelligence Today?

If you're starting today, here's how:

Look for employees who have high emotional intelligence, strong peer networks, and a genuine instinct to help others succeed. These are your glue players. They might not be your highest performers on paper. That doesn't matter.

Give them 2-3 hours per week to facilitate peer learning groups of 4-6 people. Provide basic facilitation training and visible recognition. Create a channel for them to share what they're learning back to the broader organization.

Track adoption metrics and watch how peer learning accelerates compared to formal training. Scale from there.

I am interested to see if the companies that move on this first will get a 12-18 month head start on the competition. It’s going to come down to the data for me but it is definitely a trend worth watching.

FAQ

Q: What's the difference between team intelligence and emotional intelligence?

Emotional intelligence is an individual trait. Team intelligence is the collective capacity to think, reason, and solve problems together. It emerges when psychological safety, clear communication, aligned values, and trust exist. Emotional intelligence helps individual contributors. Team intelligence helps teams outperform the sum of their parts.

Q: If AI is democratizing capability, won't collaboration naturally improve?

No. AI democratizes capability but doesn't teach trust or remove ego. Organizations with internal politics often find AI tools exacerbate the problem. If people won't share knowledge with colleagues today, they won't build on each other's AI-generated ideas tomorrow. Technology amplifies existing dynamics—good collaboration gets better, dysfunction gets worse.

Q: How do I measure team intelligence?

Team intelligence shows up in behaviors: Do people speak up with ideas without fear? Do they acknowledge mistakes quickly? Do they build on each other's contributions or compete for credit? Do they maintain cohesion under pressure? Measure psychological safety through anonymous surveys, track how often conflicting perspectives surface in meetings, and monitor retention rates for mid-to-high performers (a key indicator of team health).

Q: Can you develop glue players, or are they just born that way?

Glue players have strong emotional intelligence foundations, but the role emerges from culture. You can cultivate conditions that reward glue-player behaviors—recognition for mentorship, career paths that value connectors, leadership models that explicitly prioritize team cohesion. However, cultures that only reward individual metrics will systematically burn out people with glue-player instincts.

Q: What's the first step for a leader with a high-performer team that lacks psychological safety?

Acknowledge it first. Then examine what your promotion system, compensation structure, and recognition practices actually reward. If they reward individual achievement above collaboration, change them. Model psychological safety yourself—admit mistakes, ask for help, value input from all levels, visibly reduce status signals. Create micro-credibility by following through.

Q: How do I implement something like ServiceNow's peer influencer model?

Start small. Identify 3-5 employees who already use AI effectively and have high emotional intelligence. Give them facilitation training and 2-3 hours weekly to lead peer learning groups of 4-6 people. Provide recognition and a channel to share learnings organization-wide. Track adoption rates. Scale incrementally. The key is finding people with both capability and the instinct to help others succeed.