01 — Project Overview
Conversational
Knowledge Discovery
Philips is a global leader in healthcare technology, delivering innovative solutions across patient care, diagnostics, connected health, and clinical decision support.
As organizations increasingly adopt AI-powered digital assistants to improve employee productivity and customer support, Philips initiated the development of an enterprise AI chatbot capable of providing contextual assistance across internal knowledge bases, product documentation, and operational workflows.
"The objective was to create an intuitive conversational interface that enabled users to quickly retrieve information, interact with AI naturally, and reduce the time spent searching through extensive documentation."
The chatbot needed to seamlessly integrate with existing enterprise systems while maintaining a clean, trustworthy, and accessible user experience suitable for healthcare professionals. As part of the design team, I collaborated closely with a fellow Product Designer to refine the user interface, improve interaction patterns, iterate on key conversational screens, and build interactive prototypes that aligned with Philips' enterprise design standards.
02 — Problem Statement
The Search for
Clinical Context
Healthcare organizations maintain vast repositories of technical documentation, product manuals, clinical protocols, and operational guidelines. Accessing the right information often requires navigating multiple systems, resulting in slower decision-making and increased dependency on support teams. The existing workflow presented several challenges:
The core challenge was to design a conversational experience that simplified enterprise knowledge discovery while maintaining clarity, usability, and confidence in AI-generated responses for both novice and experienced users.
03 — Goals & Criteria
Setting
Success Targets
Design an intuitive interface for enterprise users supporting natural language dialogue.
Reduce the time required to locate relevant manuals and clinical protocols.
Increase user confidence through transparent and explainable AI interaction source links.
Deliver a scalable design system capable of supporting future AI-assisted features.
04 — Responsibilities
My Design Role
I collaborated with the Lead Product Designer throughout the design lifecycle, contributing to both visual design and interaction refinement. My responsibilities included:
05 — Interactive Prototype
Live Chatbot
RAG Simulator
Click suggested prompts below to interact with the RAG (Retrieval-Augmented Generation) search chatbot and observe the structured response layout.
06 — Design System
Reusable Component
Architecture
The project leveraged enterprise design systems while introducing reusable chatbot-specific components:
07 — Research & Discovery
Understanding
Clinical Workflows
Workflows & Patterns
- • Reviewing existing enterprise support workflows.
- • Understanding user expectations for AI-assisted search.
- • Studying complex, documentation-heavy healthcare workflows.
Integration Guidelines
- • Analyzing common enterprise chatbot interaction patterns.
- • Collaborating with Product Managers to prioritize user needs.
- • Reviewing design system guidelines to ensure consistency.
The research highlighted the importance of reducing cognitive load while maintaining transparency in AI-generated responses.
08 — Design Process
Iterative Methodology
09 — Journey Flow
Enterprise Knowledge
Discovery Journey
10 — Key Decisions & Features
Architectural Choices
Conversational First
Instead of a traditional search experience, the design prioritized natural conversations that encouraged intuitive AI interaction.
Structured Responses
Rather than long paragraphs, responses were divided into logical sections using headings, bullet points, and expandable cards.
Source Transparency
Responses surfaced supporting document links, helping healthcare professionals validate the origin of AI-generated information.
Progressive History
Previous conversations remained accessible through a collapsible history panel without cluttering the primary workspace.
13 — Technical Challenges
Navigating Constraints
Information Display
Presenting large volumes of AI-generated information without overwhelming clinical users.
Maintaining Trust
Maintaining user trust in AI-generated answers through clear explainability and source referencing designs.
14 — Final Solution
RAG Search Interface
The final solution delivered a clean, scalable conversational experience tailored for enterprise healthcare environments. By combining natural language interactions with structured response layouts and transparent information sourcing, the chatbot reduced friction in knowledge discovery while maintaining user confidence in AI-assisted workflows.
15 — Business Impact
Measured Outcomes
16-19 — Project Completion
Summary & Learnings
Working on an enterprise AI chatbot reinforced the importance of designing conversations that balance usability, transparency, and trust. Unlike consumer chatbots, enterprise users expect responses that are structured, verifiable, and directly actionable.