Feedier AI

Feedier is an AI-based Customer Intelligence SaaS platform that centralizes feedback, enriches with attributes and presents insights on consultant level reports.

Role: Consultant Product Designer

Duration: 6 months

Tools: Figma, Figjam, Figma Make, Claude

Deliverable: Design System, Redesigning all screens of their hardware software

Back to home

Designing an AI-Powered 24/7 Customer Insights Platform

This case study focuses on my work on the report builder feature of the Feedier BX Platform. However, I was responsible for the entire platform’s UI design from ideation to implementation & customer demos. I’ve simply narrowed the scope down to keep this case study focused.

NB: Feedier, while based in France, markets itself in English to expand its global reach. However its userbase & clients remain predominantly french.

The Challenge

Our reports were not actually being created BY our users but by our CS Team members..Why?

  • Most users found it too complex to add or customize charts beyond the template’s defaults.

  • When an error occurred, users had no way to identify the cause or know what to do next, other than contacting the support team.

  • Customizing the report's appearance (and other parts of the platform at the time) required HTML and CSS coding knowledge.

My Approach

Instead of starting with documentation, I focused on building a usable system immediately while ensuring it could evolve with the product.

1. Establish a Scalable Foundation

I created a variables-driven design system in Figma designed for flexibility and fast iteration.

This included:

  • Semantic color tokens designed for AA accessibility (with several AAA combinations)

  • Typography and spacing variables to enable global updates without manual component edits

  • A growing component library covering buttons, inputs, lists, dropdowns, sliders, tabs, and cards

This token-based structure meant large visual changes could be made quickly without rebuilding the interface.

  1. Stress-Testing it on a Complex Screen

Rather than designing abstract components in isolation, I began applying them directly to the Main Scan screen, the most constrained and feature-dense interface in the product.

I explored layout structures inspired by other high-density tools (such as music and video editing software) to:

  • Improve visual hierarchy and readability

  • Reduce clutter without hiding important functionality

Feedback from the software lead reinforced a key principle: the system had to support complexity rather than hide it.

3. Reframing Onboarding as a Workflow System

The setup flow was originally designed as a linear wizard, but once users reached the main interface they often needed to return to specific configuration steps depending on what they wanted to adjust.

Instead of treating onboarding as a one-time flow, I reframed it as a revisitable configuration workflow.

I explored:

  • Navigation patterns allowing users to jump back to setup steps

  • Clear communication of locked or dependent states

  • Entry points from the main screen to key configuration areas

Low-fidelity workflow scenarios in FigJam helped align stakeholders on these behaviors before committing to detailed UI.

  1. Supporting Engineering Implementation

Since the design system was intended for immediate use, I focused on practical workflows for the engineering team. The application is built using Delphi, which relies heavily on PNG assets for many UI elements. Because of this, providing fast and reliable exports was essential for implementation.

So to speed up implementation, I created:

  • Component pick-up zones showing all variants in one place

  • Fast export methods using Figma slices and copy-as-PNG shortcuts

  • Clean component structures engineers could easily reference

The goal was simple: make the system easy to use under real deadlines, not just well organized.

  1. Supporting Engineering Implementation

Later in the project, the team began considering SEMI E-95 interface guidelines, which emphasize strong visual salience and affordances in industrial software.

So to speed up implementation, I created:

  • Stronger interaction cues to increase affordance clarity

  • A more minimal visual style aligned with the product’s emerging design language

This work helped frame the design tradeoffs and provided a foundation for future compliance decisions.

Outcome

By the end of the engagement, the team had:

  • A token-based design system foundation ready to scale with the product

  • A near-complete component library capable of supporting dense scientific workflows

  • Early layout explorations proving the system could handle the product’s most complex screen

  • Workflow concepts that reframed onboarding as a revisitable configuration system

  • Practical implementation patterns that supported engineering adoption

The project gave the team a shared UI language and a scalable foundation for future interface improvements.

What we explored next

By the end of the engagement, the team had:

  • Usability validation with lab operators to test dense layout assumptions

  • Expanding the design system into interaction patterns and behavioral guidelines

  • Deeper alignment with SEMI E-95 standards and industrial UI affordances

Back to home