cresta chat agent

efficiency and clarity

while working on the chat agent, there were many things to consider. as the primary value of cresta is our agent interface, we put a lot of importance on the ui. when i joined the team at cresta, there was already a chat interface which featured chats side by side so the agents could have multiple conversations at a time and reduce the amount context shift while switching between conversations. Through my redesign of the chat agent, i designed a system which would be flexible for adding new features in the future and build a design pattern which would be easy to scale.

version 1

when i started at cresta, i began by improving the current interface, we made efforts to clean it up visually and focus all of efforts on adding table stakes features from the software the agents had been using, LiveEngage. The biggest differentiator with our product are the ai responses that we generate to improve the agent’s selling ability and conversion rates.

redesign of the chat agent

during the redesign, i wanted to make sure that the change wouldn’t be too significant from the previous interface, there were things that worked and others which needed some improvement. also knowing the agents would be spending 8 hours a day on the ui, there was a need for the visuals and interactions to be simple and focused.

focus on input

as the most important piece of the redesign, i wanted to make the chat input effective and efficient. it was necessary to focus on the different ways the input would be used and make sure that the smart features we wanted to add would be easily added without a complete redesign in the future. there were some changes we wanted to make to bring the ai responses into focus of the agents. we placed the first responses in the text box as grey help text which could be completed by the tab key. additionally we wanted to add / commands to the interface for hotkey responses and other smart actions.