Your Org Design Is Your AI Moat

Earlier this year, our CEO Mike Seckler wrote about what he calls the golden age of the small business owner. AI is lowering the barriers to entrepreneurship in ways we've never seen before. Founders can build faster, launch with smaller teams, and reach milestones that once required significantly more capital and headcount.

That's exciting. But it's only half the story.

While starting a business may be easier than ever, running one has never been more complicated.

What's changed isn't just the amount of complexity. It's when it shows up.

The operational challenges that used to emerge over years of growth now arrive almost immediately. Founders who are still defining their product and finding product-market fit are suddenly navigating payroll, benefits, compliance, tax regulations, and workforce decisions that once belonged to much larger companies.

Small businesses have always needed partners who could absorb that complexity on their behalf – not tools that helped them manage it, but partners who could deliver certainty and own the outcome. That was Justworks' founding insight over a decade ago.

That insight still powers everything we do. We are moving quickly to rewire Justworks with AI at the core, building the foundation to empower many more small businesses with even better, faster service. To do that, we have to rethink not only the technology we use, but the way work itself is organized.

Organizational design has always been a strategic growth lever, and AI raises both the stakes and the possibilities. How should a company like Justworks be structured in the age of AI? How do we make decisions closer to the work, move faster for customers, and still protect the trust and humanity that define our service?

This is our take on the org model we believe will help Justworks scale certainty for small businesses – and the questions every company should be asking as they design their future structure.

We had already been deep in the question of how to organize ourselves moving forward when Jack Dorsey and Roelof Botha published an essay in March that got a lot of attention across leadership circles.

The essay traced two thousand years of organizational design, from Roman legions to modern corporations, as one long attempt to solve a single problem – how do you coordinate thousands of people when any one person can only hold so much context? Hierarchy, they argued, was never really about authority. It was an information-routing protocol. And now, for the first time, there may be a better one.

The concepts felt very aligned with our direction of travel, giving language to ideas we were working through ourselves.

• A world model: a shared, continuously updated source of truth that every person and every AI agent in a company can read from and write to, making both humans and AI more effective.

• Directly Responsible Individuals: people who own specific outcomes end-to-end, with the authority to make decisions without waiting for information to travel up and down a chain of command.

• Player-coaches: leaders who can spend less time managing process and more time developing the judgment, craft, and capability of the people around them.

Together, those ideas helped clarify how work might move differently in an AI-enabled company. But they also surfaced a deeper open question – if AI makes software easier to replicate, and if more companies can adopt new operating models built around shared information, faster decision-making, and directly responsible owners, where does lasting differentiation come from?

We believe the answer is organizational intelligence: the accumulated record of decisions, tradeoffs, customer learnings, edge cases, values, and ways of working that represent a company-specific way of seeing the market, serving customers, and solving problems. That kind of intelligence is slow-built, path-dependent, and hard to copy.

Every company should be asking themselves: what intelligence have we actually earned? What have years of specific decisions, specific customers, and specific edge cases taught us that is genuinely hard to copy?

For a payments platform, the answer might be proprietary transaction data. For a professional services firm, it might be a methodology refined over decades of client work. Whatever the answer is, it should shape everything – what the company organizes around, what it protects, and what it asks of both its AI and its people.

For Justworks, our answer starts with a simple truth we have learned from more than a decade of working with small businesses: they need to stay in motion.

Small businesses get stuck at predictable moments. They get stuck when they add benefits for the first time, hire internationally, navigate a compliance issue, or run into a problem with an outside institution without clear guidance. In those moments, they need a partner that can own the outcome through the last mile.

This is where our organizational intelligence is unique. We have accumulated a deep understanding of the moments that create friction in the small business journey, the edge cases that make those moments difficult, and the institutional complexity behind them. We sit between small businesses and a world of government agencies, carriers, banks, and regulators that often operate in complicated, inconsistent, and not-always-modern ways. Every exception teaches us something, every edge case sharpens the system, and every customer we help through a moment of complexity deepens our understanding of how outcomes actually get delivered in the real world.

That's why we believe our future structure will be organized around two things.

First, customer movement.

Today, like many companies, much of our work is organized by domain expertise. We have teams with deep knowledge of underwriting, operations, sales, product, engineering, compliance, and customer support, and that expertise is matrixed together to deliver the service. Internally, work may move from a sales workflow to an underwriting workflow to an operations workflow. But a small business does not experience those handoffs as separate functions. They experience one journey: “I am trying to add benefits for my team, and I need this to work.”

Organizing around customer movement means shifting those teams around those lifecycle moments. You could imagine one team responsible for the full experience of a customer adding benefits for the first time – the product flows, the operational steps, the underwriting decisions, the customer communications, the service moments, and the AI agents that help move the work forward.

The goal is to bring the right expertise closer to the customer outcome, so the team can see the whole journey, understand where momentum breaks down, and improve the system faster. Over time, that tighter loop helps us accumulate more of the organizational intelligence that is unique to Justworks: where small businesses get stuck, what it takes to get them unstuck, and how to build a better path for the next customer.

Second, last-mile delivery.

If customer movement is about organizing around the moments where small businesses need to keep moving, last-mile delivery is about making sure Justworks can actually get them through those moments. In our category, many outcomes depend on third parties that small businesses do not control. AI can make more of that work faster and more repeatable, but it cannot make the outside world simple.

It also cannot remove the need for human judgment in the moments that matter most to customers. Most customer interactions are fast, simple, and self-serve. But there will always be a small number of moments where the issue is emotionally fraught or high-stakes, like when a founder gets a scary letter from a state agency or an employee is at the doctor and unsure whether their new baby is covered. In those moments, they should be able to reach a human quickly who understands their context and can help make the problem go away.

We believe there will always be a role for humans to play in supporting small businesses, especially to deliver on outcomes. For us, these are the experts who sit between our customers and the complexity around them, like customer success advocates who pick up the phone within 45 seconds when a small business calls in, benefits specialists managing relationships directly with carriers, or tax experts fulfilling obligations with state agencies.

This is an intentional design choice. We do not want our customers to become the operator of last resort when something gets complicated. And we do not want AI to create a service model where customers are faster until the moment they most need help. The goal is the opposite – to use AI to handle more of the repeatable work, so our human experts can be even more available for the moments that require context, care, and accountability.

The leaders who navigate this moment well are going to be the ones who start from their organizational intelligence: what does my company actually know, and how do I organize to make that knowledge compound?

The answer will be different for every company, but the discipline is the same. Identify the intelligence that is truly yours. Build your world model to carry it. Put directly responsible owners close enough to the work to deepen it. Make sure the exceptions and edge cases your people handle every day do not disappear into one team, one workflow, or one person’s memory, but feed back into the system.

That’s the opportunity AI creates for organizational design. Not only flatter companies, faster decisions, or more automation, but organizations that learn faster from the work that matters most.

For Justworks, that work is helping small businesses keep moving when complexity would otherwise slow them down. Our future structure is being designed around that promise – humans and AI working together to deliver certainty, absorb complexity, and make each customer experience better than the last.

Jeri Doris is the Chief People Officer of Justworks, where she leads the company’s people strategy and organizational design.