Botler.

Directing users to injury reporting and compensation information effectively by removing the in-between search steps.

Desktop chatbot

0-1 Product

Winter 2021

Botler laptop render

Overview

Botler is an AI-powered chatbot that directs users to injury reporting and compensation information effectively. Botler removes the inbetween steps involved in searching for important information like documents, forms, and more by guiding the user directly to them.

My role

Interaction Designer
Visual Designer
Prototyper

Team

1 Product Manager
1 Designer (Me!)
2 Developers
2 Content Specialists

Technologies

Figma
React
Drupal
Dialogflow

What is the WSIB?

The WSIB provides workers compensation for 5 million injured workers in 300k regulated workplaces. Our Innovation Lab of 12 works with WSIB often for new ideas and business needs. At the time, our team was working on 3 projects.

Workplace Safety Insurance Board of OntarioWSIB Innovation Lab

An internal research study ...

showed that there was a growing negative sentiment behind workers compensation after speaking with hundreds of employers and injured workers.

61% of injured workers and employers said they were unsatisfied with the WSIB websiteNew claims increased by 64% in 3 yearsWSIB website has too much infoUsers asked simple questions

WSIB's website is information rich, but not targeted.

The increase in claims over 3 years meant that many more people were going through the claims process. New injured workers and employers want to feel supported during the process and often access the WSIB website. The WSIB website is has tons of resources, but those resources are not that accessible.

WSIB Landing pageToo many links and optionsNested dropdowns, menus, linksToo many search results

An opportunity.

Speed up the information gathering process

A chatbot as a proposed solution?

Our team thought of many ideas like ...

  • Reworking the current website’s IA to prioritize key information
  • Redesign the FAQs page to be more accessible and useful
  • Improving the search page with better results

However, referencing the chatbot projects we were going to be working on when enter Q1 of 2021, I had another idea.

Chatbots can be an effective guide for usersLarger employers are less likely to use a chatbot since they know the process

Key decision.

A chatbot is our chosen solution

Constraints.

  • Must support WSIB accessibility guidelines (AAA standards and design language)
  • Must work with Drupal (WSIB’s website is built on it)
  • Complete the end-to-end design and development process in 3 weeks before we show stakeholders (WSIB Executives and Internal Team Leads)
Constraints

Early design work.

Guiding principlesComponents and variants
WireframesChatbot decision tree

High Fidelity.

Botler high fidelityAll designsUI decisions 1UI decisions 2UI decisions 3

Handoff was collaborative and feedback based.

The designs were shared early with developers to get them thinking about feasibility / implementation. The Dialogflow model was being trained and developed in parallel with design and content work. We implemented the design in 3 days (2 weeks in), allowing us to conduct a usability study for the week.

1 week user test + impact.

Key performance indicators
102 meaningful conversations55% of opens were voluntary67% of all situations were handled correctly91% satisfaction

Changes in the final version.

Final version changes

Learnings.

  • User needs + chatbots = amazing pairing
  • Guided users to helpful info while improving the understanding of the process
  • Made info on the WSIB website more accessible
  • The 3 weeks were extremely fast-paced, but it forced me to make effective decisions quickly

Next steps

Stakeholder were happy with the 1 week test and gave Botler the greenlight for far-future implementation. When Botler goes live (hopefully in 2022), measure satisfaction, user retention, and bounce rates over a longer period. Lastly, we should also look into training Botler with more complex situations.

Botler laptop render

This website was designed from scratch in Figma and implemented in Webflow by Rikin Patel in 2021.