All assigned designs implemented

Widya Robotics

Turning Behavioral Data into an Implemented AI Product Redesign

Using behavioral evidence to redesign Notulensi.id and Transkripsi.id across desktop and mobile.

More than 10 usability issues informed 108 implemented screens and more than 20 documented reusable components.

Organization: PT Widya Inovasi Indonesia / Widya Robotics
Role: UI/UX Intern
Timeline: August 2025–December 2025
Status: All assigned designs implemented
Contribution: Worked with one UI/UX staff member, two full-stack developers, and the CTO across Notulensi.id and Transkripsi.id.

01More than 10 usability issues identified
02108 desktop and mobile screens
03All assigned designs implemented

Project context

Redesigning two AI products through behavioral evidence

From August to December 2025, the work covered Notulensi.id and Transkripsi.id. Both products formed the scope of an evidence-led redesign across desktop and mobile.

The assigned output included navigation, banners, action feedback, loading states, icons, and illustrations, as well as broader product flows.

My role and collaboration

Contributing design inside a compact product team

I worked as a UI/UX Intern alongside one UI/UX staff member, two full-stack developers, and the CTO.

This collaboration connected behavioral analysis and interface design with the people responsible for product direction and implementation.

Problem

Product behavior pointed to usability issues that needed prioritization

PostHog behavioral data was used to identify more than 10 usability issues across Notulensi.id and Transkripsi.id.

The design challenge was to translate those signals into a manageable redesign scope rather than treating each observation as an isolated screen change.

Research and evidence

Turning behavioral analysis into an issue inventory

Design workspace organizing usability issues identified through behavioral analysis into grouped redesign priorities. — The synthesis translates behavioral signals into an organized issue inventory. The underlying PostHog capture is excluded because it exposes personal and session data.

From evidence to scope

Turning behavioral signals into prioritized redesign work

01
Analyze behavior
Use PostHog evidence to review how the products were being used.
02
Identify issues
Document more than 10 usability issues across the product experience.
03
Prioritize tasks
Translate the findings into 16 redesign tasks.
04
Define flows
Develop one to three flows within each assigned redesign task.

Key design areas

Improving clarity across actions, states, and product structure

Navigation

Redesign work included navigation across the product experience.

Banners

Banners were included within the assigned redesign scope.

Action feedback

Feedback states clarified the response to user actions.

Loading states

Loading states supported AI processes that require waiting time.

Icons

Icons were designed as part of the interface work.

Illustrations

Illustrations supported the redesigned product experience.

Design exploration and decisions

Comparing dashboard and history redesign decisions

Before-and-after comparisons of the Notulensi dashboard and meeting-history interfaces. — The comparison documents interface changes without claiming conversion, task-completion, retention, or satisfaction improvements.

Feedback for AI processes

Designing states for work that is not instantaneous

The redesign scope included action feedback and loading states. These states were especially relevant to AI processes that require users to wait for a result.

The work also included icons, illustrations, and Rive animations to support product feedback and visual communication.

Design system

Documenting more than 20 reusable components

More than 20 reusable components were created and documented for consistent use and developer handoff.

Motion and interface feedback

Rive animations within the product experience

Animated orange audio waveform icon created for Notulensi. — Audio waveform

Animated orange speaker icon created for Notulensi. — Audio output

Animated stack of transcript cards created for Notulensi. — Transcript cards

Animated orange loading block created for Notulensi. — Loading state

Animated document icon created for Notulensi. — Document state

Validation boundary

Behavioral findings shaped the work; post-launch metrics are not claimed

The verified evidence for this case study is the PostHog analysis, the identified usability issues, the assigned redesign scope, and implementation status.

No verified conversion, task-completion, retention, or quantitative post-launch satisfaction result is available, so none is presented here.

Implementation and project status

All assigned designs progressed into implementation

The final delivery comprised 108 desktop and mobile screens. All assigned designs were implemented.

Qualitative outcome

Clearer and more consistent design delivery

Reflection

What this work reinforced

What I learned

Behavioral signals become useful design inputs when they are translated into explicit priorities. The PostHog analysis helped move more than 10 usability issues into a structured set of 16 redesign tasks rather than a disconnected list of observations. Designing feedback and loading states also reinforced the importance of communicating progress when AI processes require waiting time.

What I would do next

The more than 20 documented reusable components created a clearer bridge to implementation. Continuing to strengthen component documentation, state coverage, and handoff context would help developers apply the system consistently as the products evolve.