CCH Axcess
CLIENT
Wolters Kluwer
PROJECT
AI Builders Comparison
ROLE
Design Director
TEAM
Builder.io, Figma Make, Lovable, v0
YEAR
February 2026
About
Goal
Original Screenshot
Home
Workflow > My View
AI Chatbot
How would I improve the Design?
Taking the exercise one step further, I explored adding several KPIs designed to better support the agent’s workflow. While I operated under the assumption that these core sections are necessary, the greater opportunity is not simply adding more data, but improving how information is structured, prioritized, and contextualized on the Home screen to support faster understanding and decision-making. The goal would be to transform the dashboard into a decision-support system that reduces friction, highlights risk, and enables agents to act immediately.
Initially, I prompted each AI tool to: “Update this tax agent dashboard to surface more high-value metrics and KPIs directly on the Home screen so agents can quickly assess status without navigating into individual sections.”
However, improving the design would require moving beyond volume of data and focusing on:
• Prioritization logic — Which KPIs are truly decision-driving versus informational?
• Progressive disclosure — What should be visible at a glance vs. expandable?
• Risk surfacing — Highlighting overdue, blocked, or high-impact items using clear visual hierarchy.
• Personalization — Tailoring metrics based on role, workload, or deadlines.
• Cognitive load management — Avoiding dashboard clutter while increasing insight density.
• Actionability — Pairing metrics with direct actions (e.g., “Review 5 at-risk returns”).
In a full-cycle project, I would validate these improvements through:
• Task analysis and workflow mapping per agent persona
• KPI prioritization workshops with stakeholders
• Usability testing focused on scan time and decision speed
• A/B testing layout density and information hierarchy
Conclusion
This exercise demonstrates how you can leverage AI as a rapid ideation partner within the Design Thinking process. In a full product lifecycle, this activity would sit within the Ideate phase. By treating AI tools as junior designers, I was able to quickly generate multiple approaches, uncover alternative interaction patterns, surface potential edge cases, and explore content and layout variations in a fraction of the time required for traditional concept generation.
For the purpose of this exercise, the source screenshot was treated as a proxy for validated user needs and prior research. In a real-world scenario, these concepts would be grounded in user insights, evaluated against business and technical constraints, and refined through usability testing and stakeholder collaboration.
