From Static Mocks to Design in Production
For years, product teams have juggled a familiar triangle: designers working in pixel-perfect tools, product managers living in specs and screenshots, and engineers shipping features in code. Users, meanwhile, only ever see the product in production. The gap between these worlds is where fidelity is lost and shipping slows down. Design in production aims to close this gap by bringing design work directly into live or production-like codebases. Instead of recreating interfaces in separate design environments, teams prototype with the same components, design systems, and contexts that power the shipped product. This shift supports more realistic code-based prototyping and reduces the translation cost between design files and implementation. As AI design tools become embedded in repositories and CI pipelines, they are turning the codebase itself into a collaborative surface—one where UX decisions, product hypotheses, and engineering constraints can be explored together without constant context switching.
Dessn’s AI Design-in-Production Model
Dessn is one of the most visible examples of this new workflow. The platform describes itself as an AI design-in-production environment that lets designers and product managers prototype inside their real codebase, without opening an IDE or running code locally. Instead of uploading a design system or exporting tokens, Dessn starts from the existing repository and wraps a design layer around it. Teams use their actual app, components, and design system as the canvas for experimentation, solving what Dessn calls the “localhost problem”: the historic need to set up a local environment just to tinker with the product. The platform provides read-only access to code and isolates projects in microVMs, emphasising security and non-intrusiveness. Early adopters, including teams at Color, Wispr, and Mercury, are already spending hours each day prototyping directly in production contexts, turning the product itself into the primary design tool.
AI as the Bridge Between Design and Engineering
AI is the underlying engine that makes design-in-production platforms more than just fancy preview tools. Large language models and code-aware systems can parse component libraries, understand design systems, and generate viable interface variations without manual wiring from engineers. Dessn’s founders argue that because LLMs are non-deterministic, each interface is not a single fixed artifact but a space of possible states. Once the platform can render components from the codebase, AI can surface many plausible product variants for teams to explore. This reframes the product development workflow: instead of meticulously redrawing screens, designers and PMs can navigate AI-generated options grounded in real code, then refine the best candidates. The result is fewer handoff cycles, less ambiguity in implementation, and faster convergence on shippable solutions. AI design tools thus act as translators between human intent, design language, and executable code.
Community, Funding, and the New Workflow Paradigm
Design-in-production is not just a tooling trend; it is evolving into a workflow paradigm backed by capital and community-building. Dessn has raised €5 million to expand its team and cultivate a global network of product builders aligned around this approach. Investors highlight that the strongest product founders tend to be highly technical and unconstrained in how they design, blurring traditional boundaries between design and engineering. Platforms like Dessn are betting that communities will form around shared practices: prototyping on real components, using AI-assisted exploration, and treating the codebase as the single source of truth for UX decisions. As these communities grow, they are likely to influence hiring profiles, team rituals, and even how companies document their product development workflows. The center of gravity shifts away from static artifacts toward living, code-based prototyping environments where experimentation feels closer to production reality.
Design-Forward Modernisation for Legacy Systems
The same design-in-production mindset is beginning to influence how enterprises modernise complex legacy systems. Traditionally, updating mainline infrastructure involved lengthy requirements documents, parallel design efforts, and risky rewrites detached from real production behaviour. Design-forward modernisation tools flip this script by placing design and UX workflows directly on top of existing code and environments. AI design tools can introspect legacy interfaces and components, then propose improved flows or UI layers that still respect underlying constraints. For product teams, this means they can explore modern experiences without rebuilding everything from scratch, validating changes in realistic production contexts. Code-based prototyping reduces the distance between proposed redesigns and actual deployment, which is especially critical in highly regulated or mission-critical systems. As more legacy-modernisation platforms adopt this approach, design in production is likely to become a default expectation, not a niche technique reserved for greenfield apps.