JPMorgan’s 318,512-Employee Problem: Jamie Dimon’s Small-Team Fix
Most businesses don’t have an AI problem. They have a “1% of everyone’s job” problem—and Jamie Dimon says that is why good ideas go to die.
Most businesses don’t have an AI problem. They have a “1% of everyone’s job” problem—and that is where good ideas go to die.
Jamie Dimon runs a bank with 318,512 employees and is still telling people the answer is smaller teams. That should make every founder, executive and middle manager with three layers of approval very uncomfortable.
JPMorgan’s real management bet is not scale
On September 12, Fortune highlighted a point Dimon made in JPMorgan Chase’s 2025 shareholder letter: the bank’s competitive battles are not won in giant divisions with glossy strategy decks. They are won by small, focused teams attacking a painfully specific customer problem.
Not “investment banking.” Healthcare pharma or medical devices.
Not “credit cards.” Chase Sapphire.
Not “small business.” Restaurateurs or law firms.
Not “payments.” Always-on digital payments with automatic currency conversion.
That is the real point. A company can be huge on the org chart and still behave like a collection of sharp little businesses at the customer level.
Dimon’s language was characteristically unsubtle. He said the teams assigned to these jobs should be small, authorised to make decisions, dedicated to the mission and able to move like Navy SEALs or Delta Force. A bit dramatic? Sure. But the operating principle is dead right.
JPMorgan reported 2025 revenue of $185.6 billion and net income of $57 billion. It also had 318,512 employees globally at year-end. This is not a bloke running a 14-person software startup from a converted warehouse telling us to “stay agile.” It is the boss of one of the world’s most regulated, complicated and systemically important companies admitting that size naturally breeds complexity, bureaucracy and complacency.
That admission is worth more than the military analogy.
Most leaders say they want speed. Then they build a process designed to prevent a junior manager from making a $500 mistake. They put six departments around an initiative, schedule a fortnightly steering committee, ask everyone to retain their day job, and act surprised when nothing meaningful ships for nine months.
That is not management. That is corporate chloroform.
The most expensive phrase in business is “we’ll add it to the list”
Dimon’s best observation is also the one most companies will ignore: when an initiative is 1% of a lot of people’s jobs, it will not get done.
Every operator knows this is true because we have all watched it happen.
Someone announces a big priority: improve onboarding, launch a new product line, fix pricing, build an AI capability, reduce churn, enter a market. Nobody is actually taken off their existing work. Sales still has its number. Product still has its roadmap. Finance still has month-end. Legal still has its queue. The new priority gets scattered across calendars like parsley on a bad pub meal.
Six months later, management asks for an update. There is a slide deck. There is a pilot. There is a working group. There may even be a logo.
There is no result.
I have lost money learning versions of this lesson. Not because the idea was stupid, but because I confused agreement with ownership. Ten people saying “yes, we should do that” is not the same as one capable person waking up every day knowing the outcome lives or dies with them.
Dedicated teams cost more up front because you must make a choice. You have to pull good people off other work. You need to name a leader. You must give them an outcome, a deadline and actual authority. And you need to tolerate the fact that they will make decisions you would not make yourself.
But the alternative is more expensive: paying smart people to attend meetings about work nobody has time to do.
Small teams are not an excuse for chaos
Here is where plenty of founders get this wrong. They hear “small autonomous teams” and translate it as “everyone can buy software, build their own process and call it innovation.” Before long, the business has 14 customer databases, five versions of the truth and an expense bill that looks like a ransom demand.
Dimon is more sophisticated than that. In the same letter, he argues that fast teams need big, common platforms beneath them: shared data, AI, coding, financial and CRM systems that can be deployed across the company.
That is the actual trick.
Centralise the boring foundations. Decentralise the customer fight.
You want one reliable system for identity, finance, data definitions, security and core customer records. You do not want each team rebuilding the plumbing because some executive thinks standardisation is oppression.
But once those rails exist, the team serving a particular customer or solving a particular problem should not need three vice presidents and a committee to change course.
This is where larger companies have an advantage if they are honest about it. Scale can fund the platforms, capital, risk controls and specialist capability that smaller competitors cannot afford. JPMorgan plainly has that advantage. But scale only remains an advantage if the people closest to the customer can use those assets quickly.
Otherwise, your size becomes a tax. Your competitors are not necessarily smarter. They are simply less burdened by internal permission slips.
The overlooked angle: the enemy is vague accountability, not hierarchy
The fashionable answer to corporate sluggishness is to flatten the organisation. Sack a layer of managers, redraw the chart, announce empowerment, serve canapés at the town hall. I have seen enough of these “transformations” to know they often just create a flatter mess.
A flat company can still be paralysed. In fact, it can be worse: nobody is sure who owns the decision, so everyone gets a vote and nobody carries the consequence.
Dimon’s model is not really anti-hierarchy. It is anti-fog.
Small mission teams need clear authority. They need a decision-maker. They need people whose jobs are materially tied to the result. They also need support from the rest of the business rather than turf wars from departments that feel threatened.
That last bit is where leadership earns its money.
If you tell a team it owns a mission but allow finance, legal, technology, risk and marketing to quietly veto it from the sidelines, you have not created autonomy. You have created theatre.
The CEO’s job is to make the trade-off visible: this mission matters; these people are responsible; the rest of us will help; and we will judge the work by a concrete outcome, not by how many workshops were held.
JPMorgan can make this argument credibly because Dimon is not pretending culture happens by accident. His letter describes constant business reviews, management meetings, deep dives, employee and client conversations, and leadership offsites. Culture, in this version, is not a values poster. It is a repeated management system.
That is less sexy than a viral memo about “radical candour.” It is also how organisations stay coherent when they get very big.
AI will expose sluggish management faster than it fixes it
Everyone is racing to bolt AI onto their business. Fair enough. There is real upside there. But AI will not rescue a company that cannot make decisions.
It will give a slow organisation faster reports about why it is slow.
The winners will be businesses that form small, high-trust teams around real commercial problems: reduce account-opening time, lift conversion in a defined customer segment, cut a manual compliance task, improve service resolution, stop leakage in a specific part of the funnel.
The losers will fund a central “AI strategy” group, appoint a committee, run 27 pilots and eventually announce they are taking a thoughtful approach.
Thoughtful is lovely. Profitable is better.
The reason JPMorgan’s approach matters is not that every company should copy a global bank. God help us if they do. It matters because Dimon has put his finger on the universal operating problem: businesses become less effective when responsibility is spread so widely that nobody has enough incentive, time or permission to finish the job.
What this means for you
If you run a business, lead a team or want to become the person who gets trusted with bigger work, do this tomorrow:
1. Pick one commercial problem, not a theme. “Improve retention” is a theme. “Cut churn among customers in their first 30 days by 20%” is a problem.
2. Name one accountable owner. Not a sponsor. Not a committee chair. One person who can say, plainly, “this is mine.”
3. Make the team genuinely dedicated. If nobody can spend at least half their week on the mission, stop pretending it is a priority.
4. Give the team decision rights in writing. Be explicit about what they can approve, spend, change and escalate. Ambiguity is bureaucracy wearing a cardigan.
5. Measure an outcome weekly. Revenue, margin, churn, cycle time, activation, errors, cash collected—something real. Do not measure activity unless activity is the result.
6. Use shared systems, not custom chaos. Let teams move quickly on customers and execution, but do not let them create another incompatible database because they could not be bothered asking.
7. Kill the initiative if it is not worth dedicated people. This is the hard one. A long list of unfunded priorities is just a polite lie told by management.
Dimon’s small-team doctrine is not revolutionary. That is precisely why it stings. We have known for years that focus, authority and accountability win. Most companies simply prefer the comforting illusion that everybody owning something means it is safe.
It does not mean it is safe. It means nobody owns it.
And in business, what nobody owns eventually becomes your competitor’s opportunity.