Managing Energy, Not Just Time: A Leadership Reality
You Are Not The Same Person You Were Eight Hours Ago I once missed a critical concession during a high-stakes regulatory negotiation at 4:15pm. A concession I would have caught in seconds at 9am. Not because the issue was buried in the documentation. Not because the counterparty had been particularly clever about it. Because I was not the same person I had been when I sat down that morning, and I had convinced myself otherwise. That cost six weeks of rework. Six weeks of calls, revised positions, internal re-briefings, and the particular exhaustion that comes not from hard work but from correctable mistakes. The kind that follows you home. What I remember most clearly is not the moment I spotted the error – that came later, in the debrief. What I remember is the certainty I felt getting on that call. Tired, yes. But experienced. Seasoned. Twenty years in the room for situations exactly like this. I had done harder things on less sleep. I would be fine. I was not fine. The Situation This was 2021. I was in a regulatory settlement negotiation – the kind where the stakes are measured not just in money but in precedent, in relationship, in what I would have to explain to a board that trusted my judgment. The counterparty’s team had pushed for a late afternoon slot. I knew this was not an accident when it happens in negotiations. Scheduling is a tactic. I knew that. I agreed anyway. By the time the call started, I had already been in five hours of prior meetings. The morning had been sharp – I had gone into an early session and caught two inconsistencies in the counterparty’s position before the first coffee had gone cold. That version of me was good. That version of me was not on the 4:15pm call. The concession slipped through in the language of an indemnity clause, framed as a minor administrative provision. In the context of the full document, in the state I was in, it read like standard boilerplate. At full capacity, the phrasing would have stopped me cold. Instead, I moved on. We concluded the call. I noted it as a productive session. This was not a productive session. The error surfaced forty-eight hours later during a legal review. What followed was not a crisis – we recovered, we corrected, we rebuilt the position. But the cost was real. Six weeks. And the harder cost: I had to sit with the knowledge that I had known, at some level, that I was not sharp enough for that call. I had chosen to get on it anyway, because the alternative felt like an admission I was not ready to make. What I Got Wrong – And Had Been Getting Wrong For Twenty Years The first thing I got wrong was the assumption that experience is a substitute for condition. For most of my career I had treated my own cognitive state as essentially stable. Adjustable by caffeine, by willpower, by the professional obligation to perform. My logic ran something like: I have navigated complex situations before, therefore I can navigate this one now, regardless of timing. This logic is seductive and it is false. Experience sharpens my tools. It does not mean I am holding them the same way at 4pm as I was at 9am. Neuroscience has been clear on this for decades – decision quality, working memory, and the ability to detect subtle inconsistencies all degrade across the day for most people, particularly after sustained cognitive load. Knowing this intellectually and actually scheduling around it are two entirely different things. The second thing I got wrong was conflating busyness with prioritisation. For years I had scheduled my hardest thinking into whatever slot remained after everything else was placed. Board preparation at 6pm. Critical document reviews at end of day. Strategic planning sessions wedged between operational calls. My reasoning was efficient: get the administrative and relational work done first, then tackle the substantive. In practice, I had it entirely backwards. I was giving the work that required the least of me my best hours, and giving the work that required everything I had the hours when I had nothing left. My calendar looked productive. My output suffered in ways I had not tracked carefully enough to notice the pattern. The third thing I got wrong was making this a personal failing rather than a structural one. After 2021, I did not immediately change how I operated. What I did first – and I say this without pride – was file the mistake under “lessons learned” and quietly resolve to be more careful next time. As if vigilance was the missing ingredient. What I actually needed was a different architecture. The sharpest executives I have worked alongside do not rely on willpower to protect their cognitive peak. They protect it structurally. They decline late afternoon calls for high-stakes decisions. They build buffers before complex work. They are, in this specific sense, harder to schedule than their less experienced peers – and that difficulty is not arrogance. It is professionalism. I have written before about the conditions under which teams perform at their genuine best – and the same principle applies at the individual level. The environment and the timing are not irrelevant context. They are part of the result. What This Means In Practice This is not an argument for becoming precious about my calendar. Most senior roles do not afford the luxury of perfect scheduling, and the executives who refuse any meeting after noon are usually protecting mediocrity more than sharpness. But there is a meaningful difference between unavoidable scheduling constraints and the habit of treating my cognitive peak as a flexible resource I can deploy wherever the day demands. The question worth asking is a simple one: do I know which hour of my day I am genuinely dangerous – the hour when my pattern recognition is fastest, my judgment is cleanest, my read
AI won’t replace wisdom
I first met her when she called market turns that saved the bank millions. She could read a balance sheet the way most people read a menu, instinctively, with taste, the numbers arranging themselves into meaning before anyone else in the room had found the right page. Thirty-one years of credit risk experience. The kind of institutional knowledge that doesn’t live in documentation, can’t be onboarded in a fortnight, and walks out the door when people like her retire, leaving quiet devastation in its wake. When we introduced AI-assisted risk tooling last year, she went quiet in every session. Not disruptive. Not vocal in her resistance. Quiet in a way that, if you weren’t paying attention, read as disengagement. I was paying attention, eventually, and what I saw wasn’t a woman failing to keep up. It was a woman calculating, with thirty-one years of precision, exactly how much it would cost her to be seen not-knowing., – Standing in the lift at Canary Wharf on a Tuesday morning in Q3, I kept returning to the session where she sat in the second row. We were running the third onboarding cohort for the new tooling, a mixed group, analysts through to senior directors. I had told myself we were being inclusive by putting everyone in the same room. When the facilitator asked participants to navigate the model interface live, I watched her pause at the screen for slightly longer than everyone else. Not long enough for anyone to notice. Long enough for me to notice. She recovered, clicked through, and said nothing for the rest of the session. Afterwards, I asked her how she found it. She said, *Fine.* And then, after a deliberate beat: *I just need to practice the file saving. The cloud thing.* The file saving. The cloud thing. This was a woman who had built risk frameworks from first principles, who had sat on credit committees shaping the bank’s exposure through two financial crises. She wasn’t struggling with the AI. She was struggling with the visibility of struggling, in front of people she had mentored, whose careers she had shaped, who still sent her questions they couldn’t answer. We had designed the onboarding for capability. We hadn’t designed it for dignity., – We got two things wrong initially. First, we assumed resistance and silence meant the same thing. They don’t. Resistance is a position. Silence is a calculation. When a senior expert goes quiet in a learning environment, they aren’t refusing to learn, they’re refusing to be seen as a beginner in a culture that has spent decades rewarding them for being advanced. These are different problems with different solutions, and conflating them wastes months. Second, we got the architecture of the room wrong. Mixed-cohort onboarding feels democratic. In practice, it creates a quiet social tax on senior participants, who must weigh the cost of every question against the impression it makes on people whose careers they influence. The learning environment we built was technically open and psychologically closed. Openness isn’t the absence of barriers; it’s the deliberate removal of the specific barriers that apply to the specific people in the room. The third thing I didn’t expect was that capability and confidence decouple under observation. She could navigate the tool. What she couldn’t do, not yet, was navigate it in public without the fluency she’d spent thirty years building in every other domain. There’s a particular kind of competence that only exists when no one is watching. Good learning design has to account for that gap, the gap between private ability and public performance, because that’s where most senior professionals quietly give up., – We rebuilt the onboarding from the structure outward. Smaller cohorts. Senior peers paired with senior peers, not because they needed protection, but because psychological safety isn’t an abstract value; it’s a specific condition created by specific design choices. We removed performance metrics for the first sixty days entirely. Progress was measured in questions asked, not tasks completed, because questions are evidence of engagement, and tasks completed can simply be evidence of avoidance. Ninety days after that Tuesday morning, she was running the internal AI literacy sessions herself. Not because we fixed her, but because we fixed the room. The capability was always there. The environment had been charging her too much to use it., – If your AI adoption numbers are disappointing, look at your learning architecture before you look at your people. The dominant assumption, that resistance to AI is about fear of replacement, technophobia, or generational lag, is wrong often enough to be dangerous. In financial services especially, where authority is built on the appearance of knowing, the real barrier is frequently the psychological cost of public inexperience. Your most experienced people are also your most exposed. They have the most to lose from being seen as a beginner, and they will quietly disengage before they’ll let that happen. The organisations that get this right don’t build learning environments that are merely open. They build environments that are specifically safe for expertise, where asking a question is evidence of intellectual seriousness, not a signal of inadequacy. That’s a design problem, not a culture problem. And design problems are solvable. AI won’t replace the woman who can read a balance sheet like a menu. But a poorly designed onboarding session will teach her that learning your tools isn’t worth what it costs, and that’s a loss no model can recover. Wisdom doesn’t need to be replaced. It needs a room where it’s safe to be new.
What Happens to Teams When You Force Them Outside Their Comfort Zone
When I Put My Team Somewhere They Had No Business Being 2021 was the year my best team almost quit. Not because the work was too hard. Because the work was the wrong kind of hard, and I had put them there deliberately, despite being told twice that I was making a mistake. That detail matters. This wasn’t a miscalculation born of ignorance. Their manager came to me with specific concerns. I listened, acknowledged his point, and changed nothing. Which is either evidence of considered leadership or a fairly confident act of institutional recklessness, depending on which week you ask me about. What happened next is the part I return to. Not the outcome, it went better than anyone, including me, expected, but the mechanics of what changed in those eight weeks, and why no training program I’ve ever seen could have produced the same result., – The Situation The context was a cross-border regulatory implementation. Five jurisdictions, compressed timelines, the usual chaos that attends anything involving multiple regulators who technically harmonized their frameworks but practically agree on very little. I made the call in the commercial negotiation. Two compliance leads, genuinely excellent at their work, technically rigorous, professionally credible, were assigned to a room where the currency wasn’t regulatory precision but commercial positioning. Different instincts required. Different language spoken. Their manager’s concern wasn’t unfounded. These were professionals whose careers had been built on structured frameworks, defined parameters, and the confidence that comes from knowing exactly what the rules say. I was handing them a situation where the rules were being written in real time by the people across the table. By week four, they were struggling. Not failing, there’s an important difference, but clearly operating in a way that cost them more energy than the same effort would have in a familiar environment. They were translating. Converting commercial signals into regulatory logic and back again, in every meeting, in every conversation. By week eight, they had stopped translating. They were reading the room directly., – What Actually Happened The first thing I noticed was how they handled ambiguity. In a regulatory environment, ambiguity is a problem to be resolved, you find the rule, apply the rule, document the application. In a commercial negotiation, ambiguity is often a tool. The other side uses it deliberately. Learning to recognize that distinction, and then to use it yourself, isn’t something you develop from a course on negotiation frameworks. They developed it by being in the room without a script long enough that improvising became the only option. The second thing, and this is the part that stayed with me, was how it changed the way they thought about risk. Compliance professionals are trained to see risk as something to be mitigated, disclosed, or avoided. Commercial negotiators see risk as something to be priced. Neither view is wrong. But someone who can hold both simultaneously is genuinely rare, genuinely valuable, and almost impossible to hire for because most hiring processes sort for one or the other. These two had been forced to carry both, in the same meeting, across eight weeks. The cross-wiring that produced wasn’t visible on any CV. But I watched it happen. The third observation, the one I keep coming back to in conversations about team development, is what happened to their confidence. Not the surface-level kind that comes from positive feedback, but the structural kind: the understanding that they could be wrong, correct in real time, and still keep moving. That specific resilience is what I mean when I say discomfort builds range rather than confidence. Confidence is a feeling. Range is a capability. You can fake the first. The second only exists if it has been tested., – What This Means I’ve written before about the connection between strategic clarity and team ownership, the idea that people move with more precision when they understand the *why* behind a decision, not just the *what*. What 2021 added to that is a harder truth: sometimes the *why* can’t be shared in advance, because the person receiving it isn’t yet equipped to trust it. You have to earn their retrospective agreement by being right. And that means accepting that you might be wrong, and that the cost of being wrong lands on them first. That’s an uncomfortable position for any leader who takes responsibility seriously. It was uncomfortable for me. The reflection I offer in *Final Reflections from the Front Lines of Finance* is that the decisions I most regret aren’t the bold ones, they’re the ones where I defaulted to safety on behalf of people who, given the choice, would have chosen the climb. The ceiling on familiar territory is real. Most people can’t see it from where they stand because they’ve never been high enough in unfamiliar territory to look back and notice the difference in the view. The implication for any organization trying to develop the next layer of leadership is this: your internal talent pipeline is being constrained not by ability but by assignment. The people who will surprise you most are probably sitting in rooms where they’re too comfortable being excellent at the thing they already know. Put them somewhere they don’t yet belong. Stay close enough to catch a fall. Far enough that the climb is genuinely theirs., – Comfort produces competence. Only discomfort produces people who eventually stop needing the net, and quietly start building one for someone else.
The Leadership Trap: Encouraging Without Holding Accountable
The Most Expensive Kindness in Leadership There is a particular kind of management failure that never shows up on a project dashboard. It does not trigger a red RAG status. Nobody logs it as a risk. It accumulates quietly, the way damp accumulates behind a wall, invisible until the structural damage is already done. We call it encouragement. And sometimes, that is exactly what it is. But sometimes it is something else: the comfortable avoidance of a necessary conversation, dressed up in the language of belief. I have been guilty of this. Most leaders I respect have been guilty of this. The instinct to protect someone’s confidence, especially someone with real potential, is not wrong. It becomes wrong when it outlasts the moment that required honesty., – The Situation Eighteen months into a critical data migration programme, this was 2022, mid-execution, the kind of phase where every slipped dependency costs you three downstream, we had a senior analyst who was, on paper, exactly what a programme manager wants. Sharp. Fast. Energetic. The kind of person who volunteers for the hard problems and makes the rest of the team feel that progress is possible. Her delivery estimates were also consistently optimistic. Consistently, measurably wrong. Not by catastrophic margins, no single miss looked disqualifying, but with a reliability that, in retrospect, was its own kind of pattern. A ten-day task would land in fourteen. A two-week dependency would take three. Each time, the explanation was credible. Each time, the team absorbed it with grace. What nobody did, and I include myself here, was name the pattern out loud. We celebrated her energy. We noted the delays privately. We did not connect the two in any conversation she was part of. By the time we did, three downstream teams had quietly stopped building her timelines into their plans. They had worked around her. Built in their own buffer, recalculated their dependencies, and said nothing, because naming it felt unkind, or risky, or above their pay grade to raise. She found out the way people always find out in these situations. Late, and from a direction she did not expect. When we finally had the direct conversation, her response stopped me. She said: *”I didn’t know it had that effect. Nobody told me.”* She was not defensive. She was genuinely surprised. And she was right to be, because we had collectively chosen comfort over clarity for eighteen months, and the cost of that choice had been charged to her account, not ours., – What This Actually Means The first insight is the uncomfortable one: the teams that worked around her were not protecting the programme. They were protecting themselves from an awkward conversation, and in doing so, they removed her ability to course-correct. When you route around someone without telling them, you are not managing risk. You are manufacturing a blindspot and handing it to a colleague as a gift. The second insight inverts what most people assume about accountability. We tend to treat correction as the opposite of belief in someone. The working assumption, rarely stated, frequently operative, is that raising a hard truth signals reduced confidence. In practice, the opposite is closer to true. Leaders who only affirm eventually build people who cannot process critical feedback when it finally arrives. And it always arrives. The longer it is deferred, the more it arrives not as a conversation but as a consequence. Accountability is not the withdrawal of belief in someone. It is the proof of it. The colleague you correct early is the one you still think can change. The one you quietly work around has already been written off, they just have not been told. The third insight is about timing, and it is the one I find most useful now. Encouragement and correction are not opposing forces on a spectrum that leaders must balance. They are the same act, performed at different moments. Telling someone they are doing excellent work and telling someone their estimation pattern is creating downstream risk are both forms of investment in that person’s future. The problem is that we have built a professional culture where the first feels natural and the second feels like a performance review conversation that needs to be scheduled, prepared for, and survived. That friction is artificial. We created it by treating directness as a special occasion rather than a baseline expectation., – What This Means for Your Organisation Most enterprise programmes have at least one version of this dynamic running silently in the background. Someone whose work is being quietly compensated for. A pattern that the immediate team has accepted as a fixed variable. A gap between what is said in one-to-ones and what is said in dependency planning sessions. The question worth sitting with is not whether this is happening on your team, it almost certainly is, but what it is costing the person at the centre of it. They are operating without accurate information about their own impact. They are being managed around rather than managed. And when the moment of reckoning arrives, as it will, they will have been deprived of the eighteen months of feedback that might have changed the outcome. The most generous thing you can do for someone with genuine potential is make the invisible visible, before the damage compounds quietly into distance., – Closing Encourage loudly. Correct early. Not because the second makes the first more credible, though it does, but because both are expressions of the same underlying belief: that the person in front of you is capable of more than the version of them you are currently protecting. The kindness that costs nothing to give is usually the kindness that costs the most to receive., –
Investing in Junior Talent: The Leadership Edge
Standing in the lift at Canary Wharf one morning, I realised the most important professional conversation I’ve had in the last twelve months happened over a flat white in a coffee shop around the corner from the office, with someone who had been on my team for eight months and whose full professional history I could not have described accurately if asked. That says something. I’m not sure it reflects well on me. Senior leaders talk a lot about inclusive culture, psychological safety, and building teams where every voice is heard. We commission surveys, review results, and nod at the right parts. Then we walk back upstairs and spend the next three weeks talking exclusively to the four people whose names appear on our recurring calendar invites. The hierarchy doesn’t disappear because we have good intentions. It calcifies quietly, one polite meeting at a time., – The Situation Last quarter, I had a catch-up with a junior analyst, no agenda, no prep, no performance review. We’d walked past each other in corridors for eight months. She was quiet in meetings, solid in output. The kind of person a busy senior leader files under “performing well, nothing to flag.” I knew almost nothing about her. What I didn’t know, and this is the part that has stayed with me, is that before joining us, she’d spent two years on the ground in East Africa building financial inclusion infrastructure. Payment rails in low-connectivity environments. Regulatory navigation across multiple sovereignties with overlapping and sometimes contradictory frameworks. The kind of operational context you simply can’t build in a classroom or a London risk team. She mentioned it the way you mention something you’ve stopped expecting people to find interesting. I sat with that for a moment. And then, because the coffee shop wasn’t a meeting room and there was no deck to move through, I asked her to tell me more. She did. And somewhere in that conversation, she flagged a gap in our risk model, a regional assumption we’d baked in without realising it was an assumption. The kind of gap that looks fine from the inside and only becomes visible when you’ve sat in the specific context it fails to account for. We reviewed it. We changed the decision. That change mattered. The conversation that prompted it had no entry on any project plan., – What the Coffee Actually Did The first thing I understood from that morning is something I should have understood earlier: proximity is not the same as understanding. I’d been physically near this person for eight months. I’d seen her work. I had a mental model of her contribution that was accurate as far as it went, which was not far at all. The corridor version of a person is a silhouette. The coffee version is a human being with a history. There’s a pattern in this that I’ve seen cause real damage to organisations. It’s not malicious. It’s the natural bureaucratic gravity of seniority, the way that as you move up, the information you receive becomes increasingly curated, filtered through layers of people who are, consciously or not, optimising for what they think you want to hear. I’ve written about how this kind of slow signal distortion quietly kills projects before anyone names the problem, the silent killer in teams is often not conflict or failure, but the steady narrowing of what gets said out loud. Coffee breaks the curation. Not entirely. Not permanently. But for forty minutes, it relaxes the hierarchy enough for people to say the true thing instead of the safe thing. The second thing: expertise does not announce itself in organisations with strong hierarchies. This is the counter-intuitive part. You might assume that genuinely valuable experience surfaces because people share it, because meritocracy works, because good organisations recognise good thinking. That’s not consistently true. What surfaces is the experience that has been given permission to surface, by title, by tenure, by proximity to the right meetings. Everything else waits. A junior analyst who spent two years building payment infrastructure in East Africa will wait a long time to be asked about it in a risk team meeting that already has an agenda. The third thing is less comfortable: the gap between what I thought I knew and what I actually knew was invisible to me until I accidentally closed it. That’s the specific failure I’ve been sitting with. It wasn’t that I chose to ignore her background. I genuinely didn’t know it existed. That’s not a knowledge gap I could have managed, because I didn’t know to look for it. The only way to find that kind of unknown unknown is to create the conditions in which people tell you things unprompted, in a setting where unprompted honesty feels safe. No survey instrument captures this. No one-to-one performance template gets you there reliably. Coffee gets you there. A walk gets you there. Lunch with no deliverable attached gets you there., – What This Means If You Lead Anything If you run a team of any size, there’s almost certainly someone three levels below you sitting on a professional history, a regional insight, or a technical understanding that’s directly relevant to a decision you’re currently making without it. Not because they’re withholding it. Because no one has asked, and the architecture of the working week doesn’t create the moment when they’d offer it unsolicited. The fix isn’t a new meeting. It isn’t an initiative. It’s the deliberate, unstructured hour, the coffee, the walk, the lunch with no outcome attached. Block it. Do it consistently. Accept that most of those hours will produce nothing immediately measurable. The one that does will make the others worth it., – I’ve been more wrong about the people closest to me professionally than I have about markets, models, or forecasts. The difference is that markets tell you when you’re wrong. People tend not to.
Why I Left a Contentious Meeting for 10 Minutes and What It Changed
The Best Decision in That Room Was Not Made in That Room A particular kind of silence descends over a contested governance meeting about forty minutes before everyone stops pretending they are listening. You can feel it. The energy shifts from engagement to endurance. People stop building on each other’s points and start defending positions they arrived with. Someone begins annotating their own slide deck. Someone else checks a phone with the careful nonchalance of a person who absolutely is not checking a phone. We were four hours into a six-hour session last quarter. Four regions. Three risk functions. One deeply contested data ownership model that had been, depending on who you asked, either almost resolved or fundamentally unworkable for the better part of eight months. The kind of meeting where the pre-reads are forty pages and everyone has read different forty pages. I have sat in enough of these rooms to know the difference between productive tension and performed disagreement. What we had, by mid-afternoon, was the latter. Then someone, someone who had barely spoken in two hours, asked for a ten-minute break., – The Moment Nobody Planned For The request landed awkwardly. We were mid-thread on a particularly thorny question about accountability boundaries, and the interruption read as avoidance. I thought, briefly, that we were about to lose momentum we had spent forty minutes building. I was wrong about what momentum looked like. We came back. Within eleven minutes, the blocker that had absorbed two months of alignment meetings dissolved. Two people had shifted positions. Not because new data had surfaced. Not because anyone had made a more compelling argument. In ninety seconds of corridor conversation, without a room watching, two people had given themselves permission to think differently. I did not fully understand what had happened until I asked one of them about it afterward. The answer was immediate and slightly disarming: *”I already knew I was going to move. I just needed somewhere to do it that wasn’t in front of everyone.”* That sentence has followed me into every complex negotiation since., – What the Room Was Actually Doing to People Here is the thing about contested enterprise decisions that rarely appears in any change management framework: the room itself is a performance space. The moment you seat twelve senior people around a table to resolve something difficult, you have created an audience. And audiences do something to human beings, they make the cost of changing your mind visible, immediate, and social. This is not weakness. It is entirely rational. In high-stakes institutional settings, the ability to hold a position under pressure signals competence. Shifting ground mid-meeting, in front of peers and stakeholders who will interact with you tomorrow and in six months, carries a real professional cost. Even when the shift is the right call. Even when, privately, you have already made it. The result is that contentious meetings frequently do not fail on logic. The logic, often, has been exhausted. They fail on the cost of being seen to change your mind in real time. The position people defend in the room is not always the position they hold. It is the position they can afford to be caught holding. The pause removed the audience. The audience, it turned out, was the problem., – On Designing What Happens Between the Agenda Items What interests me more than the psychology is the structural implication. If the break was not a rest but a mechanism, a designed permission to recalibrate outside of observation, then it can be built in deliberately, not stumbled into accidentally. I have started doing exactly that. Not breaks as courtesy, not breaks as schedule management, but breaks as architecture. In sessions where I know the contested points are coming, I plan the pause to arrive just after we have surfaced the conflict but before the positions have fully calcified. The timing matters. Too early and you have not created enough pressure to make the corridor conversation necessary. Too late and people have committed publicly to things they cannot walk back without losing face. This also means paying different attention to what happens outside the room. The corridor, the coffee queue, the thirty seconds by the door, these are not the informal margins of the meeting. In a complex negotiation, they are often the meeting. I have learned to be present in them rather than using them to check messages, which is, I realise, a more significant behavioral change than it sounds for someone whose instinct in any gap is to clear the inbox. There is something here that connects to a broader pattern I have noticed in transformation work: the most consequential moments are rarely the ones on the agenda. The conversation that unlocks eighteen months of deadlock happens in the ten minutes before the room fills. The decision that reframes a programme happens in the walk to the car park. There is an argument, and I find it genuinely compelling, that the scheduled session exists partly to create the conditions for those unscheduled ones. Which means the design of the pause is as important as the design of the agenda., – What This Means If You Lead Contested Sessions If you are running complex negotiations, across regions, across functions, across competing ownership models, the question worth asking is not only *what needs to be decided* but *where does the decision actually want to happen?* Some things can only be resolved in a room. They need witnesses, they need formality, they need the weight of a recorded agreement. But the movement toward that resolution, the actual shift in position, the private acknowledgment that the other side has a point, the quiet decision to stop performing certainty, almost never happens under observation. It happens when the audience leaves. Your job, in part, is to design those moments as intentionally as you design the agenda., – The most important conversation in a complex negotiation often happens in a corridor between two people who
The Real Signal Behind Bonus Season in Financial Services
The Sentence Before the Number Bonus season has a particular texture in financial services. There is the waiting, which everyone pretends not to be doing. There is the calibration meeting, which everyone pretends is objective. And then there is the conversation itself, twelve minutes, sometimes fifteen, in which an institution attempts to compress a human being’s entire year into a number and a handshake. Most organisations do this reasonably well. They train their managers. They prepare talking points. They ensure the number is fair, or fair enough, or at least defensible. What almost no organisation does well is the sentence that comes before the number, the one that names, specifically and without ceremony, what the person actually did and why it was hard. That gap, between the number and the sentence, is where a surprising amount of talent quietly decides to leave., – What Happened in January In January this year, I sat in a room with a team that had done something genuinely difficult. Eighteen months. A cross-regional data platform spanning four geographies, hundreds of stakeholders, and the kind of legacy infrastructure that makes perfectly sensible engineers stare at their laptops in silence. The numbers had landed. The bonuses were fair, I had fought for them to be fair, which is its own kind of exhausting process that nobody outside the conversation ever fully sees. The room was quiet in the wrong way. I had expected relief. I got politeness. And in the gap between those two things, I recognised something I should have understood years earlier: the team was not waiting for the number. They were waiting for the sentence before it. The one that said: *we know what you built, we know what it cost, and we know it would not have happened without you*. What I had prepared was thorough. What I had not prepared was precise. I had the data. I did not have the language. And watching people absorb a fair number with the affect of people receiving a utility bill, I understood, slightly too late, though still not entirely too late, that I had confused adequate compensation with actual recognition. They are not the same thing, and conflating them is one of the more quietly expensive mistakes a leader makes. We recovered. The conversations that followed were different because I made them different, more specific, more named, more willing to say the difficult thing aloud: *this was genuinely hard and you are genuinely good*. But I did not forget the quiet in that room., – What Recognition Actually Is The first thing worth saying is this: recognition is not praise. Praise is general, *you did a great job, the team was brilliant, we really appreciate everything you do*. It is the warm noise organisations make when they mean well but have not done the work of paying attention. People are polite about praise. They say thank you. They do not remember it. Recognition is specific. It names the thing. It says: the moment in October when you held the vendor negotiation together for three weeks while two of your leads were out, that was the moment. It says: the reason this platform works across Singapore and London and New York is because you made four hundred small decisions well, in sequence, without anyone asking you to. Recognition is evidence that someone was watching. That is what makes it different. Praise says *you are valued*. Recognition says *you were seen*. The distinction matters because being seen is the thing that compounds. It shapes how a person walks into the next difficult year, with energy, or merely obligation. Both will deliver. Only one will stay., – Why Institutions Default to Currency The second thing worth understanding is why most organisations reach for the number when they should be reaching for the sentence. It is not negligence, mostly. It is measurement. Organisations are extraordinarily good at quantifying what they can quantify, and a bonus is clean, it has a figure, a rationale, a market benchmark. It can be defended in a calibration meeting. It can be put in a letter. A precise sentence cannot be put in a letter. Or rather, it can, but writing it requires the manager to have been paying close enough attention to know what to write. That is the actual cost. Not the money. The attention. Most senior leaders are not inattentive people. They are overextended people who have learned to trust the systems, the frameworks, the ratings, the pay bands, to carry the weight of recognition. The systems are fine at compensation. They are structurally incapable of specificity. Nobody has ever felt deeply seen by a pay band., – The Practical Implication None of this requires a restructured compensation process, a new framework, or a leadership offsite. It requires about twenty minutes of preparation per conversation, the kind of preparation where you actually write down, before you enter the room, the two or three things that were specifically true about this person’s year. Not the general. The named. What did they do that was hard? What would not have happened without them? Where did they make a call that others would have deferred? The answers to those questions are the sentence. And the sentence, delivered before the number, changes the architecture of the conversation entirely. The number stops being the point. It becomes confirmation of something the person already knows you understand. That is not a small shift. For a team that has spent eighteen months building something genuinely difficult, it is the difference between leaving the room satisfied and leaving the room seen. And seen, as it turns out, is the thing people remember when they are deciding whether to do another eighteen months., – Bonuses compensate. Sentences remember. The organisations that understand the difference will find out, slowly and then all at once, that their people do too., –
Why I Built Sovereign AI Infrastructure for Executives
I have experienced a quiet institutional nervousness that never appears in risk registers. It surfaces in a hushed conversation after a long meeting, once the vendor has left and someone, usually the quietest person in the room, asks: *Wait, where does that data actually go?* I’ve been in those rooms. The honest truth is that for most of 2024, we weren’t asking the question early enough. The AI platforms offered to financial services executives today are genuinely impressive. The synthesis capabilities are real. The productivity gains are measurable. And the terms of service are long enough that nobody reads them in full, which I suspect is no accident. What gets buried in that length isn’t necessarily malicious, but it is consequential: the same infrastructure that makes a model smarter for you also makes it smarter about you. Those aren’t the same thing., – What we discovered when we looked closer In early 2024, we ran a pilot: a senior leadership team, a major AI platform, a strategic synthesis use case, the kind of work where you feed a model the texture of how your institution thinks: how it weighs risk, frames optionality, moves from data to conviction. The outputs were good. Genuinely good. The kind that makes you want to scale immediately, and several people in the room were ready to. I wasn’t. And I’ll be honest about why: I hadn’t read the terms carefully enough at the start. When I finally did, properly, with our legal and data governance teams in the same room, we found what we’d half-expected and half-hoped we wouldn’t. Inference rights. Model improvement clauses. Language that was technically precise and strategically vague in exactly the right ratio to be defensible but not transparent. We paused the pilot. That conversation was uncomfortable. Telling a leadership team that the tool works fine but the foundations need examining, after they’ve already started building on them, doesn’t land well. We ran seventeen additional weeks of legal and architectural review before deciding. By then, some of that initial comfort had worn off, which I count as useful., – Three shifts in how I now see this **The executives who grasped sovereignty fastest weren’t the most technical.** That surprised me, until it didn’t. The people who moved quickest were those who’d spent careers in competitive intelligence, deal structuring, proprietary research, people who understood, viscerally, that the way an institution thinks is itself a competitive asset. Information asymmetry is the moat. Once you internalise that, the idea of a third party accumulating inference data on your decision patterns isn’t a technical concern. It’s existential. **Governance deferred is liability accumulated, not risk avoided.** This is the counterintuitive truth I keep restating. There’s a tempting logic: adopt the platform now and sort governance later, when regulations clarify. In financial services, that logic is backwards. Regulations won’t clarify before they tighten. When they do, every institution that treated AI data sovereignty as a procurement footnote will spend considerable energy explaining why. Starting with governance isn’t caution that slows you down. It’s the architectural decision that makes everything else sustainable. **The real risk isn’t that the model knows how you think.** It’s that someone else does. That reframing changes the conversation. We’re comfortable with the idea that AI will learn from our data. Less comfortable, because we haven’t fully faced it, with the idea that the model improving on our data is also improving for someone else’s product roadmap, someone else’s pattern recognition, someone else’s competitive understanding of how major financial institutions make decisions. The custody banks renegotiating data licensing terms in early 2026 weren’t having a price conversation. They were having a sovereignty conversation. The fact that it happened quietly, without headlines, is itself telling., – What this means in practice We’re building infrastructure where the model learns from your decisions without learning for someone else’s benefit. Air-gapped reasoning environments. Institution-owned fine-tuning. No third-party inference trails. This isn’t a political stance or a rejection of external AI capability, we use external models extensively and will continue to. The distinction is between using a tool and feeding a tool. Between a model that improves your decisions and one that improves its understanding of your decisions on behalf of its developer. For any financial services organisation operating at scale, this is now a governance question that belongs alongside data residency, model risk management, and fiduciary responsibility. It’s not a question for the technology team alone. It’s for the board, the risk committee, and especially the executives whose strategic thinking is the asset at stake. The organisations that will manage this well are the ones asking it now, before a contract renewal, a regulatory review, or a competitor’s data breach forces the issue., – The door at the end In two years, the competitive advantage in AI won’t be which model your executives use. It will be whether the insight those executives produce belongs entirely to them.
AI Risk Control Strategy for Global Banking Operations
I first noticed something off in 2022. A tier-one bank’s AI reported zero issues across a full quarter of transaction monitoring. The risk committee reviewed the numbers, saw the clean run, and moved on. False positives were down. Processing time had dropped by sixty percent. By every measure they had, the model worked exactly as designed. A mid-level analyst thought something felt wrong. Not a hunch without reason. She had fifteen years in correspondent banking. Fifteen years watching money move through nested accounts, cross-border flows, paper entities that seemed solid one day and vanished the next. She read those rhythms like a cardiologist reads an ECG, not just the spikes, but the silences that shouldn’t be that quiet., – The Situation She flagged a counterparty. When I asked what had triggered her concern, she hesitated. The honest answer? She couldn’t fully explain it. The flows looked normal. The entity had paperwork. The model had processed and cleared it without a hitch. No single data point stood out. What she had was a shape. A pattern she’d seen before, not identical, but close enough that fifteen years of experience made her notice. The counterparty was structuring: deliberately breaking transactions into pieces to stay below automated detection limits. The method wasn’t new, but the setup, the jurisdictions, the counterparty type, the timing, fell outside the model’s training data. It had never encountered this exact arrangement, so it said nothing. Here’s what haunts me about that year. The model wasn’t broken. It did exactly what it was built to do, precisely. The risk committee wasn’t careless. They reviewed the outputs they were given. The governance process looked correct from every angle it could see. That’s the problem. The system had no visible failure mode for the people responsible for spotting failures. It had no red light. It only had the absence of one, and the organisation had, over time and without anyone saying it aloud, learned to treat that silence as safety., – What Actually Failed The first failure wasn’t the model. It was the way we thought about what the AI was doing. Some AI risk governance treats model metrics as a stand-in for real-world coverage. False positives. Processing speed. Accuracy on test data. These numbers matter. They tell you something real. But they don’t tell you what the model has never seen. A model trained on past transactions will catch patterns it knows. It won’t notice when the world changes, when a new structuring trick emerges, when a rarely used jurisdiction becomes a conduit. It can’t warn you about its own blind spots. That’s not a flaw in design. It’s how these systems learn. The second failure was subtler. When AI metrics look good for long stretches, organisations adjust their human oversight accordingly. Senior analysts spend less time on cleared transactions. Review processes thin out around the automated layer. That makes sense, you wouldn’t manually check every calculation a spreadsheet makes. But the analogy is wrong. A spreadsheet follows rules consistently. An AI model learns patterns from data, and the patterns it never saw are the ones it will never find. The oversight we’re cutting back is exactly the oversight we need to catch what the model misses. The third failure is the one that stays with me. It’s what happened to the analyst’s instinct inside the organisation before that moment. She had flagged things before that didn’t turn into confirmed issues. That’s how real pattern recognition works, not every signal leads to a finding. In some places, repeated flags without outcomes become a professional liability. Analysts learn to adjust their instincts to match what the model approves. The pressure, unspoken but real, pushes toward alignment with the machine. When the machine says nothing is wrong, insisting something is feels risky. Organisations can quietly erode that confidence over time without meaning to., – What This Means for Your Organisation If you run AI-assisted transaction monitoring, or any AI-assisted risk function in a regulated setting, the question isn’t whether your model performs well. It’s whether your oversight is built around what the model cannot see, or whether it’s been quietly reshaped around what the model can process. Those are not the same structures. The institutions getting this right aren’t choosing between AI capability and human judgment. They’re being precise about what each can actually do. AI handles volume and applies learned patterns at a scale no team can match. Experienced analysts spot anomalies outside those patterns, not because they’re better than the model, but because they’re different from it in the ways that count. The gap between the model’s world and reality isn’t something you close by improving the model. You close it by staffing it., – The best AI risk control system I’ve seen wasn’t the one with the strongest model. It was the one that knew, without doubt, where the model ended and what had to happen next. A system that understands its limits is still a system. One that doesn’t is a liability wearing impressive metrics.
Operational Resilience Beyond DORA: 2026 Perspectives
Operational Resilience Is Not a Document. It Is a Data Problem. DORA is already law. Most institutions are still working out whether they can actually comply with it. That distinction matters. There is a wide gulf between an organisation that has spent eighteen months building a compliance framework and one that can answer the questions the framework exists to answer. The first is visible. The second is rare. The gap between them is not a regulatory problem, it is an infrastructural one the industry has quietly avoided for years. The March 2026 Information Register submission is about to make that avoidance very loud. — The Room That Went Quiet Early in 2025, I sat with the CRO of a mid-sized financial institution. DORA had just become enforceable. Her team had done what looked, by any reasonable standard, like serious work. Eighteen months of it. Policy documents. Risk registers. Governance structures. Third-party mapping. They had the architecture of compliance. It was genuinely impressive, the kind of work that gets presented well in a board pack. Then someone in the room asked about the March 2026 Information Register submission. The room went quiet. Not the polite quiet of people organising their thoughts. The other kind, where everyone is doing rapid mental arithmetic and arriving at the same uncomfortable answer. The data needed to populate that register simply did not exist in a usable form. It was distributed across systems that did not talk to each other, owned by teams with different definitions of the same terms, stored in formats that made aggregation a manual project measured in weeks, not hours. Eighteen months of compliance work. The underlying data layer: untouched. I have been in enough of these rooms to know this silence is not unique to that institution. I have seen versions of it in London, in Dubai, in Mumbai. Different organisations, different regulators, identical pause. What struck me that day was not the gap itself, I had expected to find gaps. It was the specific shape of it. The team had built compliance theatre. Beautifully documented. Operationally hollow. — Three Things That Conversation Made Undeniable **The first is that most institutions confuse documentation with readiness.** This is not laziness or incompetence. It is a rational response to how compliance has historically been evaluated. Regulators asked for evidence of frameworks. Organisations produced frameworks. The feedback loop rewarded paperwork. DORA has changed the question being asked. The Information Register is not a document, it is a live query run against your actual data architecture. You cannot write your way to a passing grade. The data either exists in a coherent, mappable form, or it does not. **The second is that this is the same problem wearing different clothes.** The broken data layer that cannot support an Information Register submission is the same broken data layer that cannot support AI or ML initiatives. When organisations talk about data readiness for artificial intelligence, and the conversation comes up constantly now, they often frame it as a new investment required for new capabilities. In most cases, it is neither new nor optional. It is a pre-existing infrastructure deficit that AI ambitions have simply made impossible to defer. I wrote about the adjacent problem, the way third-party data creates hidden exposure, in an earlier piece on [rethinking third-party risk](https://lakshvaswani.com/when-partners-become-liabilities-rethinking-third-party-risk/). The pattern is the same: the problem is not the risk you can see. It is the one your data architecture cannot tell you about. **The third is the one the industry least wants to hear.** Operational resilience frameworks that rest on poor data infrastructure are not frameworks. They are documents waiting to embarrass you. The scenario matters here: a genuine operational disruption, a regulatory examination, a cyber incident requiring rapid forensics. All of them require the same thing, accurate, accessible, well-governed data about your critical functions, your dependencies, and your recovery pathways. If that data does not exist in a usable form under normal conditions, it will not materialise under pressure. I have written previously about [the human side of cyber risk](https://lakshvaswani.com/when-firewalls-fail-the-human-side-of-cyber-risk/) and the way institutional confidence about resilience tends to collapse precisely when it is tested. Data readiness is the structural version of that same overconfidence. — What This Means for the Institutions Still Building If your organisation is in the cohort that has the framework but not the data infrastructure to support it, and based on everything I am seeing, that is most of the industry, the path forward has a specific sequence. The Information Register deadline is the forcing function, but treating it as a one-time submission exercise would be a significant mistake. The register is a symptom question. The actual question it is asking is whether your data governance, your system architecture, and your critical function mapping are coherent enough to produce a reliable, auditable output on demand. Answering yes in 2026 and reverting to fragmentation in 2027 solves nothing. Continuous readiness cannot be delegated to a project team that convenes before a regulatory deadline. It has to be owned at the top, funded accordingly, and treated as a permanent operational capability, not an event. This also has direct implications for AI ambitions. Boards and executive committees increasingly want to understand when and how they can deploy AI and ML across risk and compliance functions. The honest answer is that those capabilities will perform in direct proportion to the quality of the data they run on. Closing the data preparedness gap is not a precondition for AI, it is the work itself. — The institutions that come out of the DORA era in the strongest position will not be the ones with the most sophisticated frameworks. They will be the ones that quietly fixed their data infrastructure while everyone else was still formatting governance documents. A resilience framework that relies on data you cannot actually produce is not a framework. It is a liability you have not invoiced yet.