Overview
Nectar is a CRM company that unifies social media messages into a single inbox. The existing tool couldn't keep up with the growing business needs, and users required a more efficient and standardized workflow. My primary task was to seamlessly integrate the company's latest AI feature into the platform, enhancing customer work efficiency and catalyzing product-led growth.
Result
Resolved more inquiries per day, increasing productivity by
30%
Improved user satisfaction rate by
36%
Time to complete common inquires reduced by
24%
0.1 Context
They developed the previous platform using a UI code template that lacked a user-centered design approach. Recently, with the introduction of an AI feature, the team aims to streamline the user flow and seamlessly integrate the new AI capabilities.
Upon joining the team, I conducted interviews with three small e-commerce companies—their target client demographic—and discovered some key insights
😫
Message
Overload
Managing multiple platforms often leads to losing track of messages while waiting for a single response.
😵💫
Unclear
Responsibilities
No assigned ownership for message handling
🥱
Overworked
staff
1-2 employees managing multiple platforms, repeatedly answering common questions.
0.2 Competitor Analysis for AI feature
Agent Side
Customer Side
03. Ideating and Defining
How might we improve our user experience?
Maintain awareness of other messages while processing one
Establish clear responsibility assignments and message status tracking
Implement AI chatbots and support templates to handle routine inquiries and automate repetitive tasks.
04.Solution
❓No assigned ownership for message handling
💡 Add the agent's name to clearly assign responsibility and display their status (online or offline).
We redesigned the tab to help them easily find all the messages they are responsible for. Testing showed this layout is more efficient.
In the "All Messages" tab, we added name to show who is in charge of each message.
❓ Overburdened staff responding to repetitive questions.
💡 Use AI chatbots to respond to customer messages.
Simultaneously, analyze conversation content to prioritize tasks for human representatives, ensuring a better customer experience.
AI automatically detects customer frustration through sentiment analysis and keywords, such as “answer my question”.
If urgent, AI marks the conversation “Action Required Now” to transfer to human representatives immediately. For non-urgent issues, AI marks “Waiting for Human” and continues assisting where possible.
Workflow:
1, AI automatically categorizes each incoming message (ex. product question, partnership inquiry), and responds to common questions.
2, AI flags messages needing human support as "Action Needed" or "Waiting - Non-Urgent"
3, Jess(human agent) steps in to resolve tricky conversations the AI cannot address with AI suggestions
4. Once resolved, Jess(human agent) marks messages as "Completed"
Implement AI support templates to help with repetitive tasks.
05. Wireframe
Sketch
Low-Fi
I had multiple round of discussions with the development team, and here is our final design. It preserves the existing system to the greatest extent, while the 3-column layout provides more room for future enhancements.
High-Fi
06. User Testing
Testing the Figma prototype with 5 existing users and 5 potential customers
"The AI templates are awesome; I no longer need to manually type repetitive words."
Users can finish all tasks by themselves
9/10
"The prices and return policies vary across different platforms. I hope to differentiate them more effectively"
Avg. satisfaction rate
7.5
so…
How might we differentiate different platform effectively?
💡 What about maintain the original layout of each platform.
For Email,
For post message,
