Journal / Strategy
The flat organisation is already here
There is a version of organisational flattening that most companies attempt and most get wrong. They remove a layer of management, declare themselves empowered, and then watch everything slow down as the decisions that used to flow through those managers now flow through nobody — or everybody at once.
The old model: specialists in silos
The traditional product organisation is built around specialisation. Researchers gather insights. Designers interpret them. Product managers translate them into requirements. Developers build to those requirements. Data analysts measure what happened. Each handoff introduces delay, translation loss, and a small but cumulative drift from the original customer reality.
This model made sense when the tools required genuine specialisation. Running a usability study required training. Querying a database required SQL. Building a working prototype required engineering time. Synthesising hundreds of support tickets required either a large team or weeks of manual work.
None of those constraints are as hard as they were.
AI-assisted research tools can recruit participants, run moderated sessions, and return a structured synthesis within hours. Designers can generate interactive prototypes from a brief. Developers can ask questions of their own usage data in natural language. Product managers can have a customer sentiment analysis across 3,000 support tickets ready before their next stakeholder meeting.
The specialist skills still matter — interpreting research well, designing interactions that hold up under real use, knowing what questions to ask of data. But the access barrier has collapsed. The question is whether organisations are reorganising around that fact.
What empowerment actually requires
Empowered teams — in the sense Marty Cagan describes in his writing on product organisations — require three things: context, capability, and clarity. Context means understanding the customer problem deeply. Capability means being able to act on it. Clarity means knowing what success looks like.
Most organisations invest heavily in clarity (OKRs, roadmaps, strategy documents) and lightly in the other two. Context is assumed to flow from the occasional research report or NPS dashboard. Capability is assumed to be covered by the team's existing specialisms.
AI changes what capability can mean at the team level. A cross-functional squad of three — designer, PM, developer — can now do work that previously required a support cast of researchers, analysts, and data scientists. This is not about headcount reduction. It is about reducing the distance between a question and an answer, and between an insight and an action.
Continuous customer insight as a default
One of the most significant changes AI enables is the shift from periodic to continuous customer insight.
The traditional research cycle — brief, recruit, conduct, analyse, present — runs on a cadence of weeks or months. By the time insights reach the team, the sprint has moved on. The insight becomes context rather than input.
AI-assisted analysis can run continuously in the background. Support conversations, in-app feedback, session recordings, review data, sales call transcripts — all of this can be processed and surfaced as structured insight without a researcher needing to touch it. Themes emerge in real time. Shifts in customer language are flagged automatically. Emerging friction points appear in a dashboard before they become churn.
This does not eliminate the need for qualitative research. Deep, contextual understanding of customer behaviour still requires skilled human observation. But it replaces the silence between research cycles with a continuous low-level signal that teams can act on immediately.
In the DAO and decentralised organisation literature, one recurring observation is that the organisations which scale well without heavy management layers are those where information flows freely and teams can make decisions with current data. Continuous AI-assisted insight is the practical mechanism for making that possible in a conventional product team.
Implementation: giving the team the tools
Flattening through AI empowerment is not a structural change first. It is a tooling and capability change that makes structural change viable.
In practice this means three things.
First, give designers access to research tools they can run themselves. AI-moderated user interview platforms, automated synthesis tools, and prototype testing services mean a designer no longer needs to route every validation question through a research team. They can test a concept this week, not next month.
Second, give developers and product managers access to their own data. Natural language query tools over product analytics mean that the question "why are users abandoning this flow" no longer requires a ticket to a data team. It requires a question, typed into a tool they already have open.
Third, build the habit of continuous insight review into team rituals. AI can surface the signal, but teams need to be looking at it. A weekly review of auto-synthesised customer feedback, support themes, and usage shifts becomes the briefing that informs the week's priorities — rather than a quarterly research debrief that is already out of date.
The product person's role in a flatter structure
In the DAO research and in Teal organisations, the product person's role does not disappear with flattening — it shifts. Less coordination, more strategy. Less directing execution, more defining the problem clearly enough that execution can happen without constant direction.
The same is true here. As AI closes the capability gap for individual contributors, the premium on good problem definition increases. A team that can move fast needs to be moving in the right direction. That means investing more time in understanding which customer problems are worth solving, what a meaningful improvement looks like, and how to know when you have achieved it.
The product leader in a flatter, AI-empowered organisation is less a router of information and more a setter of context. They make sure the team understands the customer reality deeply, knows what good looks like, and has the tools to find out what they do not yet know. The day-to-day decisions flow from that foundation without needing to pass through a central point.
The organisations that will move fastest
The companies that will benefit most from this shift are not necessarily the ones that restructure first. They are the ones that build a genuine culture of curiosity at the team level — where every designer, developer, and product manager is in the habit of asking questions, finding answers, and acting on what they find.
That culture is harder to build than an AI subscription. But the tools make it possible in a way it has never been before. The team that validates a concept on Tuesday, adjusts on Wednesday, and ships on Friday is not operating with more resources. It is operating with fewer barriers between a question and an answer.
That is what a flat organisation actually feels like. Not fewer boxes on the org chart — fewer steps between an insight and an action.
