Self-serve repair for millions of Canadians

I led the research and design team behind Virtual Repair, a machine-learning tool in the MyBell app that diagnoses internet issues and guides people through fixes in plain language.

The challenge

My team's portfolio of self-serve tools supported over 4.4 million customers, with clear goals: fewer technician dispatches, shorter calls and happier customers. When I joined, Virtual Repair worked in the background with no action needed from the customer. The next phase asked people to take part in troubleshooting themselves.

How might we design a tool that empowers people to resolve internet issues on their own, without contacting a live agent?

What we did

• Empathize: partnered with Fable to understand the challenges for users with disabilities, interviewed phone agents and technicians, and surveyed and interviewed customers.
• Define: studied how other self-serve platforms guide people, facilitated stakeholder workshops to define a successful support session, and audited existing content.
• Ideate and prototype: used troubleshooting data to prioritize the most critical tasks, then drafted flows from agent scripts and support pages.
• Test: ran moderated testing on UserTesting to sharpen the instructions and copy, then measured satisfaction with post-session surveys.

Outcome

Virtual Repair significantly reduced calls and technician dispatches. Exact figures are confidential, and improvements are ongoing.

Lessons learned

More projects