February’s shakeup at Tag B set the stage for a different kind of AI story: AI is not just a sidekick but the brain power behind the company, and Tag B might become its own operating system. The mood among the survivors is a mix of skepticism and curiosity. The core idea is simple: embed intelligence into the workflow so the machine coordinates reality rather than relaying messages through managers. It sounds like sci-fi with a corporate calendar, but the promise is practical: faster decisions, tighter feedback, and a culture that learns as it goes.
In a new blog post co-authored with Sequoia partner Roelof Botha, Jack Dorsey lays out a bold vision: AI will be embedded so deeply that it redefines what the company is, not just what it does. The aim is to build a continuously updated model of the business that coordinates work across functions, potentially replacing layers of traditional management. The oft-quoted line is provocative: AI doesn’t just augment the company; it reveals what the company actually is. If that sounds existential, it is—and it’s also potentially liberating for teams that want less bureaucracy and more meaning in their daily work.
AI-Driven Org: A World Model for Everyone
The core idea is a practical one: create a company world model that records decisions, progress, and problems in real time. This model, paired with a customer world model built from honest signals captured in millions of transactions, becomes the engine. The world model tracks what is happening now and what is likely to happen next, guiding both strategy and daily tasks. The AI layer is designed to compose actions from a robust set of capabilities—payments, lending, card issuance, banking, BNPL, payroll, and more—without requiring a chorus of middle managers to approve every move.
The design is fourfold. First, capabilities—the atomic financial primitives that power services rather than flashy dashboards. They must be reliable, compliant, and high-performing. Second, a dual world model: the company world model understands its own operations, while the customer world model uses transactional data to map market reality. Third, an intelligence layer that stitches capabilities into customer-ready solutions at the moment they’re needed. Fourth, interfaces—the apps, cards, and devices that deliver outcomes. The trick is that the value sits in the model and the intelligence, not just in pretty interfaces.
Block Edge: The New Roles at the Edge
To make this real, Tag B sketches three roles for the edge of the system. Individual contributors are deep specialists who receive context from the model and make decisions at their layer. Directly Responsible Individuals own cross-cutting problems and pull resources from the world model as needed. Player-coaches contribute code, build models, and mentor others, replacing the old managerial layer. The goal is to pull decision-making toward the work itself and away from glass‑box status meetings.
The transition is in early days, and Dorsey warns parts will break before they work. But the intent is clear: empower the edge to act with context, while the central model coordinates alignment and priorities. It’s a shift from telling people what to do to giving them a smart map and letting them navigate the terrain with fewer red lights and more green signals.
From Hierarchy to Intelligence: A Long View
Historically, organizations evolved through a series of constraints and solutions. The Romans built a nested hierarchy to route information efficiently, and the military later formalized span of control to keep a large army coordinated. The railroads brought similar thinking into business, followed by Taylor’s scientific management and then the matrix structures popularized by McKinsey. Each step sought better information flow, but none solved the core problem: how to coordinate thousands of independent actions without turning the organization into a bottleneck.
Tag B argues that the bottleneck now shifts from people to the system. The company world model and the customer world model become the new nervous system. When the intelligence layer detects a moment where a loan should be offered or a payment path adjusted, it composes the right combination of capabilities and surfaces it to the edge where a DR I or IC can execute. In practice, this means products and interfaces are important, but the real leverage comes from the model’s insight and the speed of its execution. It is a shift from managing work to orchestrating intelligence.
Practical Reality: Four Core Components
First, capabilities: payments, lending, card issuance, banking, BNPL, payroll—these are not mere features, but building blocks with strict reliability, regulatory compliance, and performance targets.
Second, world models: the internal model maps how the company operates; the customer model maps real-world behavior from transaction data. Together, they evolve toward causal and predictive power across markets.
Third, the intelligence layer: this is the composing engine that turns capabilities into solutions for customers at the moment they need them. A merchant’s cash flow might tighten as a seasonal dip looms; the layer could propose a short-term loan and adjust repayments automatically. A Cash App user moving to a new city could see a new direct deposit setup and a tailored card experience before they even search for one.
Fourth, interfaces: the surfaces through which the intelligence delivers outcomes. The real value lies in the world model and the intelligence behind it, not merely in the UI. When a required capability is missing, the roadmap emerges from the gap—the system tells you what to build next, not the other way around.
In this design, the organization tilts toward intelligence while keeping people at the edge where reality meets action. The edge doesn’t need a thick layer of management to coordinate it; the world model provides the context, and the people at the edge exercise judgment where human nuance matters—ethics, trust, and culturally sensitive decisions.
Is this path risky? Yes, in parts. Is it untested? Certainly. Is it exciting? Absolutely. The Tag B approach is a compelling test bed for how AI could reorganize work, accelerate learning, and redefine what a company is in 2026.
For those curious about the long arc, the article connects ancient and modern ideas about how information travels through organizations. It asks a simple question with outsized implications: what if the organization’s intelligence lives in a system rather than in people? If the answer is yes, the era of AI-driven organization design may begin in earnest—and the speed of a company could become its most valuable asset.
We’d love to hear your take: does AI-powered coordination feel like an inevitable progression or a risky reimagining of how work should get done? Share your thoughts in the comments below.
Special thanks to the original article by Jack Dorsey and Roelof Botha for the thoughtful concepts and bold ideas. Read the complete blog post here: original article. Thank you for sharing these ideas with the world.
Image prompt: A realistic, simple office scene showing a diverse team around desks with laptops and screens displaying a calm AI interface; a subtle glow hints at futuristic tech without being sci‑fi. Image filename: AI_Block_image_2026.jpg
References
- Times of India: After cutting 4,000 jobs, Block CEO Jack Dorsey outlines AI future
- Span of control — Britannica
- The matrix organization revisited — Harvard Business Review