Designing for addressing customer grievances
|Type of Project||Hackathon Project|
|Role||UI/UX Designer, Frontend Developer|
About the Project
Assist.ly was a project we came up with at a hackathon last year. We were a team of 3 members (Himanshu Shekhar and Madhurjya Pegu, being the other two), I was responsible for the design and front-end development of the product.
Assist.ly aims to reduce the problem of reputation degradation for consumer-facing organisations by monitoring Facebook posts. Using natural language processing and machine learning algorithms, the platform analyse comments on a Facebook page and classify them into different categories (payment issue, the delivery problem etc), and extract relevant entities.
These days, customers often use social media platforms like Facebook, Twitter etc. to register their complaints by directly commenting on their posts. These Facebook posts/comments which talk about their issues(often in a negative way) results in negative publicity and a cynical image of the seller.
- Want to solve issues that deserve higher priority first.
- Want to quickly solve as many issues as possible.
- Want to get help on their problem as soon as possible.
- Want to track the status of the issue.
- Want to provide feedback on the kind of help received.
We brainstormed and decided to create an AI-based platform that monitors the client’s Facebook and Twitter feed to fetch negative comments from the Facebook or Twitter page, sorts them into a category (using NLP) and assigns them a priority(based on the number of likes, views, and reactions). The higher the virality of the comment, the higher is the priority given to that issue. We also decided to give the user a unique tracking link to track the status of his/her issue.
I always create some rough sketches on paper before creating a high fidelity mockup so as to get a clear idea of what I am making. It also allows to quickly brainstorm different approaches to a single problem.
I decided to go with the 3 tabbed layout since it allows to quickly communicate with the users and you can instantly shuffle between different types of issues.
The 3-tab layout allows the staff members to quickly filter through the issues by priority as well as by the type of the issue. Additionally, I used colour to categorise the issues in different priorities. Red for high priority and yellow for low priority issues.
Once a comment is selected, the platform focusses on that particular issue, and a chat box to communicate with and update the consumer, pops up. The order ID and the priority are fixed at the top of the chat box, to allow the staff member to quickly access it while communicating with the other departments.
The support team can also delete the comment if it’s a spam message. Once the issue is solved, by simply clicking the check button, the issue moves to the archived tab and an automated reply goes to the consumer that his issue has been resolved.
Since we had only 16 hours, the design is not optimal and still needs changes, one of them being, the absence of customer feedback on the support staff dashboard. I have to still think how to allow the staff to assign the issues to particular individuals. But still things ended well & we won the first prize. 🎉