AI productivity suite - Details

Enterprise users were struggling with complex workflows, scattered data, and inefficiencies in retrieving internal knowledge or executing routine tasks. Traditional dashboards and portals were time-consuming and unintuitive, requiring multiple clicks and manual navigation.

To simplify user interaction with enterprise systems using a natural language interface, and to boost productivity through intelligent task automation and contextual assistance.

• Conducted stakeholder interviews with internal product, IT, and operations teams.
• Identified key friction points: long search cycles, repetitive queries, and tool-switching fatigue.
• Studied GPT prompt patterns and user expectations from AI chat systems.

Key Insights

• Conducted stakeholder interviews with internal product, IT, and operations teams.
• Identified key friction points: long search cycles, repetitive queries, and tool-switching fatigue.
• Studied GPT prompt patterns and user expectations from AI chat systems.

• Adopt a conversational-first approach, prioritizing clarity, trust, and control.
• Build a modular design system for consistency across tools and future scalability.
• Enable agentic workflows where users can delegate multi-step actions to the AI.
• Ensure responsiveness and accessibility across internal desktop systems.

• AI Chat Interface: Natural language interaction to search, summarize, retrieve documents, and run tasks.
• Context Retention & Memory: Persistent chat history and document linking to maintain continuity across sessions.
• Agentic Task Automation: Multi-step task delegation with AI tracking progress (e.g., “Summarize report and notify team”).
• Quick Actions Panel: UI shortcuts for frequent queries and actions powered by GPT prompts.
• Feedback Loop: Users rate responses or provide feedback, which refines future outputs.

• Developed in Figma with scalable components, iconography, color tokens, and spacing rules.
• Shared across teams for internal tools and external-facing OPEA version.
• Included variants for dark/light mode and LLM-specific interaction patterns.

• Worked closely with AI/ML engineers to fine-tune GPT prompts and response rendering.
• Paired with frontend developers to implement designs using ReactJS + Bootstrap.
• Included variants for dark/light mode and LLM-specific interaction patterns.