Case Study
Two designers delivered the output of a four-person team in four months. Agentic AI ran as the operating layer across discovery, design, prototyping, and documentation
Client
A technology provider for financial content distribution, with clients such as Wells Fargo, Merrill Lynch, BlackRock, Edward Jones, and others.
Industry
Project Duration
Problem
A financial services compliance platform inherited UX fragmentation through acquisition. The acquired products each carried their own interface patterns, design conventions, and front-end architecture.
The work was to unify the products under a single brand and build an enterprise design system that supports multiple frameworks, including Bootstrap and custom Angular.
The harder problem was structural: aligning design, engineering, and product teams around one shared system that scales as the platform grows.
Our Vision
Roko Labs used agentic AI across the design system lifecycle, from discovery to documentation, with Cursor, Claude, Codex, and Atlassian integrations via MCP (Model Context Protocol).
Agentic AI ran as a continuous system, not a point solution. It supported discovery, design, prototyping, and documentation. AI workflows processed large datasets in real time, mapping patterns at a scale manual review could not match.
We validated outputs with internal stakeholders and key customer accounts, aligning the system with both technical constraints and go-to-market priorities.
Our Approach
Interconnected AI workflows interacted with project artifacts, documentation, and communication platforms in real time. The stack: Cursor, Claude Code, Codex, and Atlassian (Jira and Confluence), connected through MCP (Model Context Protocol).
The AI systems shared project context rather than operating in isolation. Outputs stayed consistent across discovery, design, and delivery. The design team prototyped engineering-ready designs. AI managed alignment across the project's constraints.
The result was an AI-native operating layer that let a small team move faster, stay coordinated, and take on more work
Process
AI ran as the coordination layer across every phase: discovery, design, prototyping, documentation, and engineering requirements.
Key Outcomes
2-3x Team Leverage.
Two designers delivered the output of a four-person team in four months by using AI across discovery, design, and documentation.
Engineering-Ready Design.
Designs were built inside real frameworks such as Bootstrap, not abstract mockups. Handoff friction dropped and rework fell.
Faster Delivery.
AI cut time spent on analysis, iteration, and documentation. Timelines shortened, quality held.
Automated Workflows.
Documentation, reporting, and backlog creation ran automatically. The team focused on review and judgment.
Context-Aware Knowledge.
AI delivered real-time access to accessibility standards, framework conventions, and engineering norms, with shared context running through MCP.
Stronger Alignment.
Shared documentation and visibility improved decisions across design, engineering, and product.
2–3x
Team Output. A team of two designers delivered the equivalent output of a four-person team.
4 Months
Accelerated Delivery Timeline. Designers were able to deliver 2X the work within a four-month period.
~50%
Smaller team than traditional engagements.
Near-Zero
Design-to-Engineering Rework.
