I posted something on LinkedIn a while back that got 165 reactions, which for my audience of financial services professionals is a lot. The topic wasn't AI. It wasn't banking. It wasn't Silicon Valley vs Europe.
It was about the brain.
Specifically: the fact that every time you switch between tasks, your brain takes approximately 25 minutes to get back to the same level of concentration. Not 25 seconds. Twenty-five minutes. That's research from the University of California, not a motivational poster.
The reaction told me something. Everyone knows multitasking is bad. Nobody acts like it.
The Tax Nobody Calculates
Think about what a typical day looks like for a relationship manager at a private bank. They're reviewing a client portfolio, then a Slack message pulls them into a compliance question, then an email from a prospect, then a meeting about a different client, then back to the first portfolio. Each switch is a 25-minute cognitive reset.
If you switch tasks just four times in a morning — and most people switch far more often than that — you've lost nearly two hours of deep work. Not to meetings. Not to breaks. To the invisible friction of context switching.
Nobody puts "lost 2 hours to context switching" in their time tracker. But it happens every single day.
Andrea Giuliodori, in his book Riconquista il tuo tempo, lays this out with a table that should terrify anyone who manages knowledge workers. The more tasks you juggle simultaneously, the more time evaporates into the void of switching. It's not linear. It's exponential. Three parallel tasks don't cost 3x — they cost 5x or more in effective productivity.
Why This Matters for AI
Here's where this connects to what I do every day.
One of the most powerful applications of AI in financial services isn't some fancy autonomous agent or predictive model. It's reducing context switches.
An AI that prepares a client brief before a meeting means the banker doesn't have to pull data from four different systems. That's not just saving 20 minutes of manual work. It's saving the cognitive switching cost of opening each system, remembering where they left off, finding the relevant data, and mentally reconstructing the client's situation.
The actual time saved by AI in a workflow like this is far greater than the task itself. Because the real cost was never the task. It was the switching.
Most banks that buy AI tools measure ROI in time-per-task. They should be measuring it in context-switches-eliminated. The second number is always bigger.
The brain is not designed for multitasking. It is designed for focus. Every tool, every process, every AI integration should be evaluated by a simple question: does this reduce or increase the number of times someone has to switch context?
The Open-Office, Always-On Trap
Modern work environments are designed to maximize interruptions. Open offices. Slack channels. Email notifications. Calendar invites. Each one is a context switch waiting to happen.
I work between Milan and Menlo Park, and the pattern is the same in both places. The people who produce the most meaningful work — whether it's closing a deal, building a product, or writing a strategy — are the ones who protect their focus blocks like appointments.
In Silicon Valley, this is becoming common. Engineers at serious companies block their calendars. "Focus time" is a calendar event. It's respected, or at least acknowledged.
In European financial services, the culture is still overwhelmingly reactive. Being "always available" is treated as a virtue. Responding to emails within minutes is expected. Taking two hours to think without interruption is seen as being disconnected.
This is backwards. The most valuable thing a senior professional can do is think deeply about one thing for a sustained period. Everything that prevents that — including well-intentioned Slack messages and "quick syncs" — is destroying value.
What I've Changed
I'm not claiming expertise here. I'm sharing what works for me after years of getting this wrong.
Morning blocks. I don't check email or Slack before 10am. My mornings are for the hardest thing on my plate that day — usually a proposal, a strategy document, or preparation for a critical meeting. Two hours of uninterrupted work in the morning produces more than six hours of interrupted work in the afternoon.
Batching communications. I respond to messages in two or three batches per day, not in real-time. Nobody has ever complained. Most "urgent" messages aren't urgent — they feel urgent because someone else is in reactive mode.
One screen, one task. When I'm working on something that requires thought, I close every browser tab, every app, every notification channel except the one thing I'm doing. It sounds extreme. It's not. It's baseline hygiene for knowledge work, and we've normalized the opposite.
AI as a switching reducer. I use AI tools specifically to avoid context switches. Instead of opening four tabs to research a prospect before a call, I have an AI summarize everything into one brief. Instead of switching between CRM, email, and notes to prepare for a meeting, I have it all synthesized. The goal isn't to do tasks faster. It's to do fewer switches.
The Math Nobody Does
If the average knowledge worker switches context 300 times a day — which is a conservative estimate based on research from RescueTime — and each switch costs even 5 minutes of reduced effectiveness (far less than the 25-minute full-reset), that's 1,500 minutes lost per day across a team of ten people. Twenty-five hours. More than three full working days, every single day, evaporating into the friction of switching.
No company would accept losing three employees' worth of productivity to a process they could see. But because context switching is invisible, it gets ignored.
The irony is that most productivity tools — including most AI tools — actually increase switching. They add another tab, another notification, another system to check. A tool that saves five minutes per task but adds two more context switches per day is a net negative.
The Simple Rule
The brain is not a browser. It doesn't have tabs. It has a single thread of deep attention, and everything you do either protects that thread or fragments it.
Every meeting that could have been an email is a context switch. Every notification that doesn't require immediate action is a context switch. Every "quick question" in Slack is a context switch for someone.
The most productive people I know, in Silicon Valley and in European finance, are not the ones who do the most things. They're the ones who switch the least.
That's what my LinkedIn post was about. That's what 165 reactions told me people already know but haven't operationalized.
The question isn't whether multitasking is bad. The research has been clear on that for twenty years. The question is whether you're willing to structure your day, your team, and your tools around the way the brain actually works — or whether you'll keep pretending it has infinite tabs.
It doesn't. And every tab you open has a cost you can't see.