⬅️  Back to /ClientWork

Sinitic

SaaS Design

Sinitic is a pioneer in multilingual customer experience technology in Asia. Using natural language processing (NLP), they enable chatbots to converse in local languages almost naturally. Their products are used by stakeholders like the Central Bank of the Philippines.

Deliverables

Link to client website

Link to Figma design

Brief

Sinitic, now Proto, is the pioneer in multilingual customer experience in Asia, using natural language processing (NLP) technology to create chatbots that understand local languages. Their products are used by stakeholders like the Central Bank of the Philippines, administered by the RegTech for Regulators Accelerator and Rockefeller Philanthropy, and sponsored by Bill and Melinda Gates Foundation.

As it entered a new stage of growth, Sinitic was looking to redesign their platform. Their legacy application was functional, at best, and was failing several new clients requirements. castlelab worked closely with Sinitic CTO Albert Zhuang, and his team of developers, on designing a first-of-its-kind, multilingual chatbot and financial complaints processing system.

Role

As one of my longest projects to date, the nature of my involvement varied throughout. I was the Account Lead; I sold the work, scoped it, and onboarded all team members. I was also the Project Manager and Lead Designer responsible for our day-to-day deliverables.

Team

  • Albert Zhuang, Product Owner (client team)
  • Two contractors, UI Designers
  • Leo Chazalon, Lead Designer

SNAPSHOT OF 
DELIVERABLES 

Use the Figma integration to browser through the hundreds of elements, screens and interactions designed for this project. Some of the following screens were white-labelled to preserve the confidentiality of client information.

From April to July, both teams worked hand-in-hand through rounds of user feedback to ultimately launch Sinitic's flagship product: a customer complaints processing system. The information architecture follows the the user journey in customer support:

  • LAUNCH - set up the chatbot framework and import historic data.
  • EDIT - build the chatbot dialogue tree to service common complain scenarios
  • TRAIN - using machine learning, help your chatbot improve by labelling correct responses
  • CHAT - take over the automated conversation, and classify conversations for Customer Support managers
  • TRACK - data about chatbot and team performance

LAUNCH

create chatbot and define language
overview of chatbots and their language training

EDIT

Building a chatbot model adapted to the nature of a business and the associated customer support scenarios.

Blank canvas, and test chat
JSON API Bloc configuration
Carousel Bloc configuration
Add languages to existing chatbot

CHAT

An interface for customer support agents to take control over automated conversations, and for CS managers to classify conversations and forward them to corresponding teams. 

chat history
messaging interface - CS agent view

TRACK

open tickets - manager view

ANALYTICS

customer support performance

PERMISSIONS

platform user permission

WEBCHAT

Sinitic webchat embedded on customer website