Cisco AI Assistant for Webex
Cisco AI Assistant for Webex Control Hub helps transforming the ITs work by conversational AI, scouring massive amounts of product documentation and help articles, synthesizing analytics reports, and even proactively troubleshooting.
This product launched in October 2023 for its initial version and continues to enhance. Check out the feature blog here. The demo shared in 2023 year-end WebexOne conference.
I led the design from define the key user scenarios to interactive flow. Collaborated closely with product manager, visual team, and engineer group. Please contact me for the details.
Title
Lead Product Designer
Year
2023 - 2024
What problems we are trying to solve in Control Hub?
Synthesize data
"How do I see energy use of my Webex devices?"
Configure settings
"Allow the name labels in face recognition."
Generate reports
"Download the last 7 days meeting active hosts report."
Troubleshooting
"Show me the issue of Donna Miller. Can I set up alert for her?"
First impression
Conversation prompts
The AI Assistant targets users with questions while also offering starter pointsto guide discussions, encouraging engagement, and fostering inclusivity, particularly for early-stage.
Communication tone
Clarity: The platform's tone should focus on clear communication to ensure users easily understand responses and instructions.
Professionalism: Maintain a professional, formal, and respectful tone, reflecting seriousness and credibility.
Helpfulness: Be empathetic and proactive, providing assistance, guidance, and solutions as users encounter challenges.
How to define a good experience in conversation?
Design principles
Direct Manipulation
Allow users to control the conversation flow, providing options for navigation, clarification, or correction as needed.
Progressive Disclosure
Provide information in a progressive manner, revealing more details as needed and avoiding overwhelming users with unnecessary information upfront.
One task for one time
Limiting questions to one at a time enhances the user experience by preventing information overload and ensuring that users receive relevant and actionable responses.
Follow up
Context Awareness
Design AI to understand and remember context from previous interactions, allowing for more personalized and relevant responses.
Actionable
Including action items below the information ensures its practical utility.
Error Handling
Handle errors gracefully, providing clear and actionable error messages that help users understand and resolve issues without frustration.
Takeaways
I have the privilege of working on an avant-garde AI product that transforming the work style of IT admins.
The technical black box forms the backbone of its features, emphasizing that design value extends beyond visual aesthetics to encompass the entire user experience, fostering better collaboration with product managers and engineers.
Establishing trust in AI is challenging; hence, transparency regarding data usage permissions and privacy is crucial.
To manage expectations, it's essential to communicate the learning curve for the AI system, signaling ongoing improvement efforts.
Conversational AI UX challenges
• Building a feature that people don’t know what they expect
• Communicate design values in an engineer-driven environment
• Prove the concept from demo to implementation
Other projects you may want to check
Create components, variants, inclusive design, and principles
Unfortunately due to NDA, I can't share too much detail about this project.
However, I'm happy to offer a general overview of AI assistant design. If you're interested in delving deeper into my design, please feel free to reach out.
Conversation modeling
A new format of Information Architecture