FUSE

broken UX compounded with unvalidated AI

Reported October 2026

What was the error?

While considering switching to Fuse Energy, I encountered four failures in the customer journey:

1.Poor website UX: The tariff box wasn’t clickable. The only option was to “Start switch”, preventing me from easily viewing the tariff details.

2.Incorrect AI guidance: The chatbot confidently gave me instructions that didn’t work, directing me towards functionality that wasn’t available as described.

3. No direct human handover: The chatbot couldn’t transfer me to a person, forcing me to waste time composing and sending a separate email.

4. Failure to listen: The human response answered the immediate tariff question but failed to address my wider concern about the poor experience and the need to learn from it.

Individually, each failure was avoidable. Together, they turned a prospective customer into a lost customer.

What would have been a better experience?

Better UX: Make tariff rates and charges accessible before customers commit to switching.

Validated AI: Ensure chatbot instructions are tested against the actual website and are accurate.

Seamless escalation: Enable direct handover to a human when AI cannot resolve an issue.

Listen and learn: Address the customer’s wider concern, acknowledge the failure and use the feedback to improve.

The key lesson for ErrorX

The most significant failure was not any individual error, but the apparent absence of a feedback loop to ensure those errors are fixed for the next customer.

The chatbot didn’t resolve the problem or facilitate a human handover. The human response failed to address the wider concern. And nowhere in the experience was there any indication that the feedback would be captured, analysed, assigned to an owner or used to improve the customer journey.

The learning risks disappearing into a black hole. The next customer could encounter exactly the same failures.

This is precisely what ErrorX aims to fix: capturing failures wherever they occur, turning individual customer experiences into actionable intelligence, and ensuring lessons are owned, acted upon and embedded into the system.