July 16, 2026

The Architecture of Trust

Teamship in the Spotlight, A Cross-Domain Analysis by Jenson8. What ten conversations reveal about the hidden logic of teams.

Over the past year, Jenson8 has sat down with ten outstanding individuals, across sport, the military, banking, human rights, crisis leadership, technology, and global development, to ask a deceptively simple question: how does teamship actually work, and does it work the same way in every environment? The series is called Teamship in the Spotlight, and the conversations span Rachel Malcolm (Scotland rugby captain), Andy Carre (former Army officer who commanded soldiers in Basra), David Eldon (former HSBC Chairman in Asia), Leslee Udwin (filmmaker and human rights activist behind India’s Daughter), Naila Chowdhury (crisis leader who has run organisations ofup to 17,000 people through conflict zones), Serena Gonzales (AI and talent executive), Ilian Mihov (economist and former Dean of INSEAD), Jackie Baresfield (Learning & Development leader at BAT), and Shomy Hasan Chowdhury (co-founder of the youth leadership organisation Awareness 360). What follows draws out the patterns that emerged once their answers were placed side by side.

Why teamship resists measurement, and why that is the point

Economics spent most of a century modelling humans as rational agents, utility-maximisers who weigh costs and benefits and act accordingly. Professor Ilian Mihov,macro economist and former Dean of INSEAD, built his early career inside that model. Then he became a leader of people, and the model broke.

"People do not experience a decision purely as a rational outcome," he found, working through diagnostic exercises with his own executive committee ; "they experience it through trust, fear, status, and belonging." His analytical, direct, low-tolerance-for-repetition style , the very traits that had made him an effective economist- was read by colleagues as dismissive. Logic was necessary but not sufficient. Teamship, he concluded, is "the ability to listen to others, respect their viewpoints, agree on an objective and be willing to live with disagreement while still moving forward together", a definition built not on optimisation, but onthe deliberate management of an unsolvable problem.

This is the thread that runs beneath all ten conversations. None of these leaders frames teamship as a soft add-on to performance. Nearly all of them frame it as the load-bearing structure beneath performance, the thing that determines whether intelligence, technical skill, and strategy actually convert into results under pressure.

What follows is not a list of nice quotes. It is an attempt to extract the under lying mechanisms these leaders are independently describing, because when a rugby captain, an economist, and a war-zone administrator arrive at structurally identical conclusions from entirely different starting points,that convergence is itself data.

Finding one: Trust is not a feeling; it is infrastructure

Every interviewee described trust in functional, almost engineering terms, not as warmth, but as the mechanism that allows a group to act as a single unit under uncertainty.

Andy Carre, who commanded soldiers under mortar and small-arms fire in Basra, separates trust into two distinct load paths: lateral trust between peers ("confidence that peers will do what they say, support you and tellyou the truth") and vertical trust up and down a chain of command. At Sandhurst, 25–30 total strangers had roughly five weeks to build both before the remaining 44 weeks of training depended on it. "Trust is the glue that brings everyone together," he says , not decoration, structural adhesive.

Jackie, a Learning & Development leader at BAT, arrived at a near-identical definition through a completely different route: the modern"opportunity market place" model of work, where cross-functional teams assemble around skills rather than job titles and might get "one kick-off meeting and a Teams link" instead of months to bond. Her diagnostic forwhether that speed-built trust is real is what she calls the crunch moment test: the instant someone could quietly protect themselves, by blaming a colleague, taking credit, omitting an awkward truth, and you watch what they do."Team badges and values posters fall away. What remains is character." This reframes trust not as an emotional state that precedes action, but as a pattern of choices that is only verifiable retrospectively,under exactly the conditions that make it costly to honour.

Naila Chowdhury, who led organisations of up to 17,000 people through conflict zones, makes the same claim in the language of interdependence rather than testing: "Without trust, I could not have gone one step further." For her, the causality runs in a direction most leadership literature ignores; trust is not something a leader distributes downward, it is something built mutually, in both directions, over time: "On the journey,they built me too." In some countries, that trust had to be built before her teams could function at all: certain cultural norms meant women were not permitted to work with her organisation until their own families had come to trust it, which meant, for Naila, that trust-building was never a soft preliminary to the real work; it was the precondition for having a workforce.

David Eldon, former HSBC Chairman in Asia, arrived at the same infrastructure through a very different route: his mother’s three rules, given to him as a boy whose family worked "the back stairs" of wealthy households, be loyal and honest; never ask someone to do something you would not do yourself; and always explain your reasoning when you ask for something. That third rule, he says, is where most leaders fail: they issue instructions without the backstory and later wonder why compliance feels hollow. He tested for trustworthy team players with a strange but effective ritual, throwing aball at a job candidate mid-interview. Whether they caught it and threw it back, rather than flinching, told him more about their instinct to engage as ateam player than any answer on a CV.

The convergence across combat, corporate restructuring, banking, and social-impact leadership is precise: trust functions as infrastructure that must be built before load is applied, verified only when load is applied, and built bidirectionally regardless of hierarchy, and, per Eldon and Chowdhury,it often has to be earned from people who have no formal reason to extend it at all.

Finding two: Psychological safety is a crisis behaviour, not a comfort setting

The single most consistent finding across all ten conversations concerns what psychological safety is, and it is not what most corporate training implies.

Rachel Malcolm, Scotland’s rugby captain, gave the sharpest possible version: "How people behave in tough times is the biggest indicator of whether a team is safe or not. Teams that turn on each other when things go wrong and pass blame are unsafe places." Safety, in her formulation, cannot be assessed during calm periods at all, it is a variable that only becomes visible under stress, which means most organisations are measuring the wrong moments. Her diagnostic test for it is disarmingly simple: "When someone makes a mistake, and nobody panics. Nobody blames. We reset and go again." On her worst day on the pitch, she says, "You give what you have, even if it’s 60%. The team absorbs the rest", a description of psychological safety not as comfort, but as a load-sharing agreement that only exists because it has been tested and honoured before.

Jackie’s version locates the same mechanism inside routine corporate life: high-performing teams "feel like this on a Monday morning: you walk in with energy because you’re free to focus on your strengths, not on politics or self-protection." The absence of anxiety is a downstream effect of a deeper condition, that failure will be met with correction, not blame. She adds a structural correction to how psychological safety is usually discussed: it is not a perk a leader grants, it is a distributed responsibility that every team member either builds or erodes through the choice to speak up in the room versus save their real views for corridor chats afterwards.

Shomy Hasan Chowdhury, co-founder of the youth leadership organisation Awareness 360, introduces a variable most safety frameworks miss entirely: the asymmetric cost of honesty. Speaking up about a decision, a value conflict, oran uncomfortable truth "can risk someone’s job, visa status or ability to support their family." Psychological safety, in her framing, is not equally available to everyone in the room; it is a privilege distributed unevenly according to how much a person has to lose. "It’s a privilege tobe able to grieve, to process our emotions," she notes; for many people, survival calculus overrides authenticity long before any leadership behaviour comes into play.

This is a genuinely under explored dimension: most psychological safety literature assumes a level playing field of risk. Shomy’s observation, echo edobliquely by Naila’s descriptions of women in some environments needing to stay silent even during emergency drills, suggests safety cannot be built through leader behaviour alone; it requires actively correcting for who bears disproportionate risk in speaking at all.

Serena adds a further layer that most safety frameworks also miss: the vocabulary of distress is not universal. In some cultures, she notes, a stressed colleague will say "I’m feeling overwhelmed"; in others, the same emotion surfaces as "there’s a lot on my plate at the moment "the feeling externalised onto circumstances rather than claimed as an internal state. Neither is honest. The danger, she argues, is a leader trained to listen only for the first register, mistaking the second for stoicism, and missing the signal entirely. "I think that comes down to the leader to be able to interpret these cultural signals." Safety, in other words, is not just unevenly distributed by risk; it is unevenly detected, depending on whether a leader has learned to hear more than one dialect of struggle.

Finding three: The unnamed emotion problem

Ask people what makes teams function, and they will describe processes, communication cadences, feedback loops, and role clarity. Ask what holds teams together in extremis, and multiple interviewees, independently, reached for a word rarely used in business.

Andy Carre’s answer, when asked what emotion teams feel but rarely name, was blunt: "love." In Basra, s urrounded by mortar and small-arms firefrom 360 degrees, what held the team together was "deep mutual care, born of shared experience, training and an unspoken commitment not to let each other down." He is explicit that this is not a battlefield-specific phenomenon: "that emotional bond is as relevant in boardrooms as it is on the battlefield."

Naila Chowdhury’s parallel is recognition, "the unnamed emotion teams crave and rarely get." The loudest voices in an organisation are routinely celebrated while its quietest, hardest-working people go unseen andl eft unaddressed; that gap breeds resentment even when performance metrics look fine. Her own antidote was to make herself the first to break the silence: she held weekly sessions where she asked her team to tell her everything, she had done wrong, and cried in front of them, telling them plainly, "Without you, I would have collapsed." That admission of need, rather than undermining her authority, was what gave others permission to bring their own struggles into the open. Serena, describing bias in AI and talent environments, points to a related gap: teams suppress their quieter strengths and avoid conflict specifically because the emotional cost of surfacing disagreement has never been named or normalised.

Leslee Udwin, the filmmaker-turned-activist behind India’s Daughter,supplies the sharpest possible distinction for why naming the emotion matters at all. Most organisations, when they address this territory, reach for perspective-taking, the familiar instruction to walk a mile in someone else’s shoes. Leslee is blunt that this is a cognitive exercise, and can be performed at arm’s length, even cynically. Empathy, by contrast, is affective: "we feel another’s joy or suffering sufficiently that our own concerns temporarily recede." The distinction is not academic. "Collaboration without empathy can be exploitative; conflict resolution without empathy can be mere damage control; critical thinking without empathy can become sophisticated rationalisation of injustice." In other words, most of the capabilities organisations already claim to value are hollow without the one emotion they rarely name.

What is notable is not that these are different emotions, care,recognition, safety to disagree, empathy, but that they share a structural feature: each is systematically under-discussed relative to its actualload-bearing function, precisely because business language has no established vocabulary for it. Ilian Mihov makes this explicit from the outside, as a researcher: genuine empathy is "one of the emotions leaders struggle most to show, even when they feel it," because it risks being perceived as weakness, bias, or a threat to an image of calm control, even though"thoughtfully expressed emotion can strengthen a leader’s credibility" by signalling that a leader sees human stakes, not just the spreadsheet.

The pattern suggests teamship has an em otional substrate that organisations have systematically under-instrumented, not because it doesn’tmatter, but because naming it has historically been read as unprofessional.

Finding four: Toxicity and bias are not culture problems, they are trust-destruction mechanisms

Several interviewees were asked what single behaviour they would remove from teams everywhere. The answers cluster around two mechanisms that function almost identically despite looking different on the surface: toxicity and bias.

Andy Carre is categorical: "You cannot be an effective leader andtoxic at the same time." He notes that in commercial settings, toxic high performers sometimes survive because short-term results mask the damage, butthe cost is "seismic" and unavoidable eventually. Critically, he has never seen anyone successfully manage their own toxic tendencies; his conclusion is structural, not developmental: identify and remove at the selection stage, because remediation does not work.

Serena’s parallel diagnosis, from AI and talent leadership, is that"biases are the one thing that takes down every single good teamthing." She describes co-facilitating a major industry consortium, on stage alongside two older colleagues, despite holding a doctorate and the title of Chief Talent Officer, and being positioned in the feedback afterwards as"an assistant." Attendee feedback forms independently flagged that the room had been dominated by men and Americans, and one attendee even askedwhether an AI could analyse who had spoken the most. When Serena surfaced thatpattern honestly, she says, the men in the room grew uncomfortable, and the discomfort was then quietly redirected back onto her for raising it: "This is why women never speak up, because when they do, rather than reassuring them,the men feel uncomfortable, and it’s held against the women for being uncomfortable." That single moment of bias functioned structurally exactly like Andy’s toxic individual: it didn’t just cause offence; it actively degraded the group’s ability to use the talent in the room, then punished the person who named it .

Serena extends the same diagnosis to the tools organisations now trust to be neutral: "our human biases are influencing AI, you just ask it to draw the perfect woman, and there’s your bias," pointing out that the same distortion appears if you ask for "the perfect man." Left unchecked, she argues, artificial intelligence does not correct human bias; it laundersit, reflecting "a very broken society" with an authority that makes the bias harder, not easier, to see.

Leslee Udwin’s work is the theoretical anchor for why this happens. After 31 hours of interviewing men responsible for extreme violence, her conclusionwas that violence "is not the disease, it is a symptom" of a deeper mindset that ranks some lives, contributions, or voices as worth less than others. Toxicity and bias are not separate problems from different domains; they are the same underlying mechanism (a hierarchy of whose voice counts) expressing itself at different intensities, from a dismissive comment in a meeting to catastrophic violence.

This reframes a common organisational assumption. Toxicity and bias areusually treated as HR or culture issues, addressed through policy. The interviews suggest they are better understood as trust-destruction mechanisms with measurable second-order effects on team cognition, because a team that has learned some voices don’t count has, functionally, reduced its own intelligence.

Finding five: Leadershipis a distribution problem, not a casting problem

Rachel Malcolm names the myth directly and rejects it in the same breath she would reject being called the sole custodian of culture: "If I’m the only one driving standards, we’re in trouble." She deliberately deploys "lieutenants" in high-pressure moments, specific teammates whose strengths complement her own, and is explicit that leadership becomes dictatorship, and teamship collapses, the moment it centralises. Her closing line for the entire conversation makes the distribution argument in a single image: "Real teamship begins when you care more about the badge on the front than the name on the back."

Andy Carre’s version inverts the organisational chart entirely: servant leadership means "the leader sits at the bottom, setting boundaries, resources and the sand box in which the team operates," while the team is given room to grow, take ownership, and contribute inside that structure. Thisis not laissez-faire; it is active condition-setting that deliberately distributes agency rather than command.

David Eldon’s banking-world version is a direct warning against the opposite instinct: leaders who hire clones because it is easier. He would"much rather have a team that argues," because disagreement surfaces the information that a single perspective cannot see. He borrows Charles Handy’s image for what happens when organisations fail at this: a warehouse divided into separate rooms, each division operating as if the others do not exist. His practical antidote was almost aggressively unglamorous: monthly, informal, cross-divisional coffee catchups, with no agenda beyond making sure people from different rooms knew one another before a crisis forced them to collaborate. Naila’s servant-leadership framing, explicitly credited to Ken Blanchard, who told her she was "the role model" of the concept he had written about, insists that "no individual can take the credit,"because a team’s output is not attributable to any single node.

Serena’s formulation is the most concise articulation of why this matters mechanically: leadership in a functioning team flows to "the most empathetic person in the room… and that might or might not be the leader."This is a testable, almost engineering claim, that authority and functional leadership are distinct variables, and teams that conflate them (by assumingthe person with the title is also the person best positioned to lead in each moment) are running on a design flaw.

Finding six: AI does not replace teamship, it raises the price of getting it wrong

Every interviewee who addressed AI arrived at a structurally similar position, despite radically different vantage points.

Ilian Mihov, whose own discipline is now being reshaped by AI-assisted modelling, draws the boundary precisely: AI can "accelerate analysis,scenario generation, and even drafting of strategic options, but it cannot take responsibility for the decisions those options support." He extends this into a genuinely novel argument grounded in neuroscience and physics: human brains perform equivalent cognitive functions at a fraction of the energy cost of current AI architectures, "if our brains had similar energy consumption to what AI uses for similar cognitive functions, our brains would need to be thousands of times larger” and physical constraints such as the speed of light place hard limits on distributed computation. His conclusion is not AI-scepticism; it is a claim that we should expect powerful AI, but not magical AI, and that human judgment remains irreducible precisely because responsibility, unlike analysis, cannot be distributed to a system that bears no consequences.

David Eldon’s illustrative case makes this concrete rather than theoretical: a factory loan application that looked perfect on paper, solid accounts, strong order book, that any automated system would likely approve inminutes. Walking the factory floor himself, he was reading signals no spreadsheet could capture, whether the plant was orderly, whether staff seemed engaged, whether the operation felt well-run. Those human-only signals would have over ridden the numbers had they contradicted them.

Shomy’s contribution names a subtler risk: the availability of AI tools makes it easier for people to disguise not-knowing as expertise, quietly asking a chatbot and presenting the output as their own thinking, eroding the very norm she considers essential to healthy teams, the ability to say "I don’tknow" without penalty. Leslee Udwin’s concern operates at the societal scale but shares the same logic: "our human biases are influencing AI," meaning a technology trained on a biased society will, without active counter-pressure, reproduce and potentially amplify precisely the trust-destruction mechanisms named in Finding Four.

The composite finding: AI does not reduce the importance of teamship. It removes the friction that used to force human deliberation, meaning teams can now move faster toward a wrong, biased, or unaccountable decision than at any point in organisational history, which makes the human capacities described throughout this piece (trust, distributed leadership, named emotion, resistanceto bias) more valuable, not less, exactly when it becomes easiest to skip them.

The synthesis: Teamship as a decision-making architecture under irreducible uncertainty

Taken individually, these are ten interesting stories. Taken together, they describe something more specific: a common architecture for how groups make good decisions when no formula exists to guarantee one.

Ilian Mihov’s line, arrived at through economics and organisational psychology, could stand as the thesis for the entire series: teamship is "less about finding solutions than about learning to live with problems, managing them with and through others rather than hoping to eliminate them." Rachel Malcolm arrives at the same place through sport: teams don’t win by eliminating pressure, they win by how they behave inside it. Naila Chowdhury arrives at it through crisis leadership: trust is not a precondition secured before hard moments arrive; it is what is actively built during them. Shomy arrives at it through a global non-profit spanning 70 countries: belonging and accountability are not opposites to be traded off, but two conditions that must coexist for honesty to be possible at all.

None of these leaders describes teamship as an emotional nicety layered on top of "real" performance metrics. Every one of them describes itas the mechanism by which uncertainty, disagreement, bias, and risk get processed into a decision the group can stand behind, under fire, under a P&L, under a deadline, or under the eyes of 500 young people from 70 countries.

Two closing interviews capture the whole argument better than any framework could. During the 2012 Olympic opening ceremony rehearsals, a sodden car parkin Dagenham standing in for an Olympic stadium, Serena recalls a director telling a crowd of exhausted, unpaid volunteers: "Everything we do here is much, much bigger than you, and it is much, much bigger than me. But you knowwhat? It’s not bigger than all of us together." Naila Chowdhury, asked to complete the sentence "real teamship begins when…", didn’t pause:"Trust, trust, trust. Nothing can go forward without it. And it has to show up in your behaviour, your work and your management, every day. "Between a car park in Essex and a war zone in South Asia, the same structureholds: something has to be bigger than any one person in the room, and trust is the only mechanism that lets a group act on that belief before it has been proven true.

The uncomfortable implication for organisations is this: if teamship is genuinely adecision-making architecture rather than a cultural preference, then failing to build it deliberately is not a values gap. It is a capability gap, as seriousas a gap in financial controls or technical infrastructure.

At Jenson8, this is precisely why we build immersive, experience-ledenvironments rather than relying on workshops and slide decks: because everyleader in this series arrived at the same conclusion through direct, oftenpainful experience — teamship cannot be taught as a concept. It must beencountered, tested, and practised under conditions that resemble the pressureit is designed to survive.