Role

UX Designer

Duration

2 Months

The AskWSIB Chatbot project was initiated to address a critical challenge faced by WSIB Case Managers and RTW (Return-to-Work) Specialists. These professionals rely on the Operational Policy Manual (OPM) to guide decisions regarding workplace injury cases. However, accessing this information in the field, often from PDF documents on mobile devices, was cumbersome and time-consuming.

Problem Space

WSIB staff rely on the Operational Policy Manual, or the OPM, to make informed decisions about benefits, recovery and RTW plans. This helps ensure injured workers receive fair and timely support.


But currently, accessing and navigating the OPM is a cumbersome process. Workers must either sift through a web page with several nested documents or, during fieldwork in remote areas, consult PDFs stored on their phones or paper manuals.

Empathize

Understanding the User

To gain a deep understanding of the challenges faced by Case Managers and RTW Specialists, we connected with various stakeholders and explored existing workflows. Our research involved engaging with various chatbot teams within WSIB and participating in a general learning session with an external organization. These sessions provided critical insights into the gaps and opportunities in chatbot design, as well as the broader context of user needs. We also synthesized research from user experience articles and past WSIB projects on Generative AI and human support chatbots.


Example interview Questions

  • How can we ensure the chatbot is intuitive and accessible, even for non-technical users?

  • What strategies can minimize error rates and frustration in conversational flows?

  • How do we balance chatbot capabilities with the need to maintain professional expertise (avoiding deskilling)?


Key Insights from Research

  • Users need quick access to accurate information to make decisions in real-time.

  • Many operate in the field, without reliable internet access.

  • Navigating PDFs was inefficient and detracted from their primary goal—supporting workers.

Define

Our mission was to design a chatbot that would provide these professionals with immediate, contextually relevant answers to their queries, streamlining their workflows. This project also served as an exploration of AI-driven solutions within WSIB, aiming to demonstrate how emerging technology could meet internal needs and, in the future, expand to customer-facing applications. We decided to come up with a list of our project constraints to guide or design.

Project Constraints

Standalone Mobile App

The chatbot had to function as a self-contained tool, accessible without external systems.

Offline Functionality

Ensuring usability in remote areas without consistent internet connectivity was essential.

Small Language Model

The chatbot would use a SLM, limiting advanced features typical of large-scale AI models.

These constraints shaped the design strategy, requiring careful prioritization of features and user needs.

Design and Iteration

We began designing based on our initial external research of best practices, and our conversations with WSIB chatbot designers. Our main focus was creating an intuitive and accessible interface that aligns with WSIB’s design standards. The chatbot was designed as a standalone mobile application capable of functioning offline to meet the needs of Case Managers and RTW Specialists in the field. We began by designing the core functions of onboarding and messaging, and moved on to the built in manual from there.

Onboarding

Messaging

Side Menu and OPM

Key Features

Response Format

Before

After

We experimented with different response formats. Initially, we designed the response to appear in a message bubble, resembling a typical text conversation. However, we found that removing the bubble maximized screen real estate, reducing the need for excessive scrolling – a critical consideration for mobile devices with limited space. Additionally, we incorporated formatting techniques, like numbers and bolding, to organize content and enhance readability.

Error States and Edge Cases

In the event of potential issues or unexpected situations, we proactively prepared for 4 common scenarios, ensuring the chatbot remains friendly and provides actionable solutions:

  • If the user asks a question that is relevant, but outside of AskWSIB’s capabilities, AskWSIB recommends related topics.

  • If the user asks a question that is completely outside of AskWSIB’s capabilities, The bot directs users to WSIB's website.

  • In the case of technical issues, AskWSIB guides users to the feedback form or IT support.

  • If users submit an incomplete question, AskWSIB prompts users to clarify their query.

Chat Interactions

To make the chat more practical and user friendly, we decided to input copy and edit features. These simple features improve upon minor inconveniences such as the user needing to type in the same question in a different way. By allowing them to edit their message, they can save time and interact with the chatbot in a more effective way.

Feedback and Iteration

Our two month timeline didn't allow for us to conduct proper user testing, so this meant that we had to constantly test, iterate and repeat. Our main resource for feedback was weekly design critique sessions hosted by the WSIB's internal design team. We received plenty of feedback and made minor adjustments through the app. One main area that I was in charge of re-designing was the built in Operational Policy Manual. The initial design was based off a combination of external design research, and the current WSIB web version which ensured that we were aligned with WSIB design standards.

Initial Design

OPM Home

OPM Subsection

Individual Pages

The main areas I was tasked with redesigning were based off these two feedback points.


  1. Clearer separation of subheadings in the menu list.

  • Members of the design critique commented on the list view being a bit confusing. They mentioned that they wished it was easier to see which subsection was selected or opened. They also encouraged us to try different methods of indicating separation rather than bolding/colour.


  1. More intuitive navigation

  • A hamburger menu within a chatbot usually opens up chat history, allowing the user to click on any given chat when they want to reference it. Peopled mentioned that the hamburger menu inside the individual pages of manual was confusing and they weren't sure where it would lead to.

Final Design

OPM Home

OPM Subsection

Individual Pages

I addressed the issues around clearer separation by making researched back iterations.


  • Instead of using contrasting colours to indicate when a section was selected, I inputted chevrons to provide more responsive feedback to the user. Next I made sure that the main titles remained black but each subsection was blue and indented to provide further clarity to the user.

  • By improving both responsiveness and visual clarity within the OPM, the user has a easier time navigating through the sections.


To solve the navigation issues, I made sure that the nav elements were intuitive.


  • I left the hamburger menu on the OPM home page because this is the first screen the user would see when coming from the chatbot section. Since they had to interact with the hamburger menu to open the OPM section of the app, it is best for the user if they can navigate back to chatbot in the same way.

  • For the individual pages of the application, I elected to remove the side menu. The back arrow alone was enough for the user to know how to get back to the home page, and having the option to see chat history through the side menu would just confuse them more.

Reflection

The WSIB chatbot project provided a valuable opportunity to explore how conversational AI can enhance the accessibility and efficiency of WSIB services. By focusing on user-centric design and aligning the chatbot’s capabilities with organizational needs, we aimed to create a tool that improves both internal workflows and client interactions.


If we had more time, conducting more iterative user testing would have allowed us to refine the conversational flow further and ensure a seamless experience across diverse user groups. Exploring multi-language support and adding detailed analytics for performance tracking could also have elevated the project's impact.


On a personal level, the WSIB chatbot project was an invaluable experience that not only allowed me to contribute to improving WSIB's client services but also taught me how to thrive in a fast-paced environment. Balancing tight deadlines and dynamic requirements pushed me to prioritize tasks efficiently and adapt to changing project needs.


This project reinforced my ability to work collaboratively under pressure while staying focused on delivering user-centered solutions, and I’m proud of the progress we achieved within the time constraints.