Why New Technologies Don’t Transform Incumbents

The challenge of AI adoption is proving frustrating for many incumbent companies. They’re deploying the technology widely—adding it to existing workflows, using it to automate individual tasks, incorporating it in new productivity tools—but nonetheless they’re losing out to smaller challengers. This is a familiar pattern. Promising technologies emerge, incumbents invest heavily and articulate well-reasoned strategies, yet they still end up with little more than incremental efficiency gains, and as a result they cede advantage to smaller challengers. When incumbents flounder in this manner, their leaders tend to blame execution. They point to familiar challenges, citing skills gaps, training that lags technological change, and middle management inertia. But that’s all a convenient diagnosis, not an explanation. The problem lies elsewhere: Incumbents fail because they haven’t recognized that new technologies enable new ways of coordinating work. AI, for example, allows knowledge work to be broken into smaller pieces, completed more quickly, and recombined in new ways. When incumbents struggle to leverage AI effectively, they are typically using its capabilities merely to streamline existing roles and processes. They’re doing the same work a little more efficiently. But the bigger opportunity lies elsewhere. Companies can also use those capabilities to restructure the organization of work—how it’s divided, sequenced, and coordinated across teams. When companies do that, some strategies simply stop being feasible, no matter how well-thought-out and achievable they might appear. But at the same time new strategies become possible, and those often are the source of competitive advantage. Figma vs. Adobe The rise of the collaborative-design platform Figma during the past decade shows how a new coordination architecture can determine strategic outcomes long before product features or execution matter. Adobe, the incumbent in the design space, is often held up as a model of successful digital transformation. In the early 2010s, as cloud-based technologies were emerging, Adobe abandoned its boxed software model and made a successful transition to a subscription model. By every conventional benchmark, the company behaved exactly as a forward-thinking incumbent is supposed to. And yet during this same period Figma, then a tiny competitor, built a shared design environment that reshaped how teams collaborated and made decisions—one that Adobe could not match. That ultimately created a strategic threat so great that Adobe would make a $20 billion acquisition attempt in 2022. (The deal was later terminated after more than a year of scrutiny by regulators in Europe and the UK.) Observers usually attribute Figma’s success to a superior user experience, incorporating real-time collaboration, browser-based access, and faster user onboarding. But Adobe itself built many of the same features. The real source of Figma’s advantage—and of the divergence between the two companies—was in how each understood the nature of design work and organized itself around that understanding. Adobe was locked into an older model, built on the assumption that designers edited files individually, handed them off, and integrated changes later. In this model, the cloud provided storage and distribution benefits, but file-based work and sequential editing persisted, carrying assumptions that shaped workflows and team structures in the company’s customer organizations. Coordination required managing and reconciling changes across different file versions. Figma, unencumbered by these constraints, reimagined a new way of coordinating design work. Leveraging the cloud, it allowed multiple people to simultaneously work on the same design project in a shared environment. Instead of waiting for file handoffs, teams using Figma could see one another’s changes as they happened and converge on decisions faster. Adobe treated the file as the basic unit of work, but Figma decided to treat the elements inside the file—text blocks, styles, reusable components—as the basic units of work: the smallest pieces that could be edited and managed independently. One person could rewrite text, another could change fonts and colors, a third could rearrange screens, and a fourth could connect the screens into a working flow, all at once, in the same design, without creating conflicting versions. Because each element was uniquely identifiable, a change made once automatically would update everywhere it was used, for everyone working on the design. Most importantly, treating elements as the unit of work removed a key bottleneck of file-based tools. Design work flows through many teams, and managing that work through files creates friction every time changes have to be reconciled or reapplied. Figma entirely eliminated this problem by managing design elements in shared libraries that teams across the organization draw from. Because text, styles, and reusable components are treated as shared reusable elements rather than tied to individual files, their properties can be defined once and reused across the organization. A change to a font, color, or layout rule made by the team that owns it automatically shows up in every design that depends on it, without anyone needing to open, copy, or update files manually. The change also had an important effect on organization-wide governance. Design standards at Figma are built in by default rather than enforced through after-the-fact reviews. Teams don’t have to constantly check whether work complies with guidelines—the work is built using components that already do. Because teams assemble designs using shared components, consistency emerges naturally even when teams don’t explicitly coordinate through reviews and reconciliation. By the time the market recognized that Figma had brought about a new way of organizing and governing design work, the difference between the two companies was encoded deep in their systems. Adobe was committed to an outdated, file-centric architecture, and Figma was not, which helped Figma gain market share, while also transforming the nature of design work. The Shein Advantage The companies that truly benefit from AI do not merely overlay it on top of legacy systems. Consider the case of the Chinese retailer Shein. On the surface, Shein appears to be a fast-fashion company competing on low-cost production. But its real advantage lies in how it reorganizes the unit of work in fashion—and, with it, the structure of coordination, the capabilities that become accessible, and the competitive strategies that follow. Traditional fashion houses still operate around the fashion season as the unit of work. Design, forecasting, buying, merchandising, and production are all coordinated around a long-cycle calendar. Even when these houses adopt generative AI to speed up design sketches, their underlying structure of work remains sequential. Shein, in contrast, reduces the unit of work from the season to continuous in-market experiments with new product concepts. Every design brief introduces a new product concept that can be tested in the market. Instead of trying to project customer demand months in advance, the company captures social-media trends and engagement metrics and generates new design briefs instantly on that basis. What’s really changed at Shein, though, is the way that work is organized. Algorithmic systems replace manual buying decisions, triggering small production runs for each product concept, which allows Shein to run thousands of tests in parallel in the market. A specific product concept is scaled with further batches only when demand is confirmed using actual market data on the test run. All of this becomes possible because the unit of work changes from the seasonal collection to a testable product concept. Fashion incumbents, stuck to seasonal calendars, rely on multiple layers of approval to coordinate work across a long production cycle. Shein, instead, uses AI across trend detection, design, and production, such that all functions are coordinated in response to changes in demand. The work of design and production is no longer sequential but operates in a continuous loop of sensing, generating, testing, and refining fashion concepts. In Figma’s case, changing the unit of work transformed organization-wide governance, enforced through shared components rather than through oversight. Shein’s shift goes further to create a new learning system. Because each product concept is small, testable, and independently scalable, every market interaction generates usable feedback. And because coordination is continuous rather than calendar-based, that feedback is absorbed immediately into the next round of design and production decisions. AI doesn’t sit on top of this process as an optimization layer; it serves as the central mechanism that connects sensing, decision-making, and action into a single learning cycle. The organization learns through constant, low-cost experimentation during its operations rather than through periodic reviews. Over time, this continuous learning compounds into an advantage incumbents struggle to match, not because they lack data or models but because their units of work and coordination structures prevent effective learning. New Workflows When a new workflow is established inside an organization, it gradually restructures the organization around it. New coordination patterns then determine where teams need to be positioned relative to one another, and which roles gain or lose influence. Over time, those adjustments accumulate into a new organizational architecture that reflects the structure of the work more accurately. As this restructuring unfolds, two dimensions tend to shift together. One is the unit of work: the basic object of production, whether a file or a component. The other is the coordination mechanism, which determines whether work moves through sequential approvals or is synchronized continuously. Incumbents typically change along only one of these dimensions at a time: They adopt new tools while keeping the old unit of work intact, or they deconstruct tasks while preserving sequential coordination. Challengers, by contrast, move diagonally across both: They redefine the unit of work and reorganize coordination around it simultaneously, creating workflows, roles, and decision structures that incumbents cannot easily replicate. That change in how work is coordinated reshapes how teams operate, which then shapes the products a company builds and the ecosystem that forms around them. When leaders see disruption coming yet struggle to respond, often that’s because they’re operating inside organizational architectures built for a different way of working. The Way Forward The real difference between Figma, Shein, and their incumbents is not vision but bottlenecks. Figma removed the file as a bottleneck to design decisions; Shein removed the season as a bottleneck to learning what works in the market. Incumbents keep those bottlenecks in place, which means that they can’t pursue strategies that require fast iteration, continuous learning, or fine-grained coordination. So even when leaders articulate bold strategies, the organization can only pursue those that fit through its narrowest bottlenecks. The question executives must ask today is not whether they’re adopting AI aggressively enough but whether the basic structure of work inside their organization still makes sense, given what AI now makes possible. If the unit of work remains too large, too bundled, or too slow to change, no amount of tooling will produce more than incremental gains. The problem is that units of work define roles, expertise, and status inside organizations. Changing them redistributes influence—away from those who manage approvals and toward those who design new ways of organizing work. Leaders who sense resistance often mislabel it as cultural inertia when in fact it amounts to architectural self-preservation.

Why New Technologies Don’t Transform Incumbents

Summary.

Partner Center