Wano
UI Designer
2023
Wano was an AI platform that allowed businesses to train and launch their AI Agents. It was focused on creating AI employees to assistance their human counterpart. The most fleshed version of this was the Customer care AI, that was designed to engage with customers with trained knowledge of the business.
I was the founding the UI designer on a team of 2 designers.
The first phase had many iterations and many moving parts that had to be designed out the focus was for the UI to be easy as possible but even with that, we included guided tours for new users to the dashboard.

Memory
The Memory page is where business owners teach their AI assistant about their company. We split the knowledge base into simple categories About, FAQs, Product/Services, Legal, Support, and Other so users always know what to upload and what's still missing, with required sections clearly marked.
The upload experience was kept deliberately simple: drag and drop, or upload from your device, with a template option for anyone starting from scratch. Below it we added a sources table to show everything the assistant has learned so far, giving business owners full visibility and control over their AI employee's knowledge.

Users could immediately test their customer service ai after uploading their information to the Agents memory, this was designed so businesses knew what their customers would be interacting with. The UI designed to be a small modal similar to what would be on their websites.

Instructions
This page is where business owners shape how their AI assistant talks to customers. I broke it into categories: Requests, Enquiries, Complaints, and Scheduled Instructions so owners can set specific guidance for each type of interaction rather than one long confusing rulebook, this was also down to prevent any form of hallucination on the agents side
For anyone unsure where to start we included a templates section that sits right at the top.

Conversations
This is where business teams monitor and step into live customer conversations handled by the AI assistant. Every chat is listed with its status at a glance active, awaiting handoff, assigned, or resolved so teams can quickly spot which conversations need human attention.
Inside a conversation an auto generated summary captures the key context and any data collected so a human agent never has to read through a long chat history to get up to speed. When a customer requests a real person, the assistant flags it clearly, and a single Takeover button lets the agent step in seamlessly without breaking the conversation flow

Widgets
We designed the widget to be clean and clear to understand. Although it looked like other support widget we decided to lean into what users were already familiar with to reduce friction when using the widgets.








