joanarbo.com

What It Is

Joan Arbó is a Madrid-based fractional Design System Architect. The site is not positioned as a general-purpose design tool, but as an architectural consulting service for enterprise-grade design systems. Its core argument is that, in an era where AI Agents can compose, orchestrate, and extend interfaces, enterprise design systems should not focus only on visual consistency—they also need machine-readable, testable, and maintainable system structures.

Core Capabilities

The service methodology covers five layers. First, it audits and consolidates existing systems, identifying redundancy and compressing them into a smaller, more robust component foundation. Second, it treats design tokens as the API between design and engineering, establishing semantic contracts. Third, it designs components as modules that can be discovered, composed, and orchestrated by AI Agents. Fourth, it replaces static PDF documentation with Storybook, visual testing, and API contracts, emphasizing “documentation as code.” Finally, at the end of a three-month engagement, it transfers tools, knowledge, and ownership to the client team.

Pricing and Delivery

The website does not disclose specific pricing or payment methods. What is clear is that the service follows a fractional project model: only two to three projects are accepted each year, each lasting three months. It explicitly does not offer indefinite retainers and does not encourage scope creep. This model suits teams with a defined budget, clear goals, and a desire to complete an architectural upgrade quickly, but it is not a good fit for clients needing long-term on-site support or low-cost execution outsourcing.

Pros and Cons

Its strengths lie in a highly focused positioning, covering tokens, components, documentation, testing, deployment validation, and team handover, with large-scale experience at Amazon Alexa, Western Union, and others as supporting credibility. The site also showcases five case studies across education, real estate intelligence, wedding planning, regulatory intelligence, and health apps, suggesting that the methodology can be transferred across domains. The limitations are that the public information leans heavily toward brand storytelling and lacks details on pricing, contracts, copyright, specific tech stacks, and compatibility with Figma or frontend frameworks. Limited availability also means higher uncertainty around procurement and scheduling.

Best Fit and Access from China

This is better suited to mid-to-large enterprises with an existing product portfolio, cross-market operations, or significant design system debt—especially teams preparing to involve AI Agents in interface generation and operations. For users in China, access status cannot be determined from the available text; payment options, Chinese-language support, and localized delivery are also not specified. If a local alternative is needed, teams could consider domestic DesignOps consulting, internal frontend architecture teams, or building their own governance system based on ecosystems such as Ant Design, Semi Design, or Arco Design.