Project Brief

One of our large energy clients had been working to integrate AI into their company to improve processes and efficiency. They created their own AI chatbot that pulled content from their vast collection of data sources. The goal of the chatbot was to allow employees to rapidly gain answers to compliance, legal, and regulatory questions and to search through contracts.

Expert Review Yielded Actionable Direction

My company was contracted to help them improve the overall user experience of the chatbot in order to increase user adoption. I started by conducting an Expert Review of the chatbot. For each issue I found, I explained the issue, why it was a problem, and showed how it could be improved or fixed. Each issue was then assigned a severity rating. I delivered the Expert Review via a remote workshop and was able to provide the client with actionable steps that they immediately put into play.

Some of the issues I found:

• Response time was too slow and the UI lacked any indicators that the system was actively processing the request
• The prompt-area buttons were not obvious in their functions, which could result in users clicking every button to see which one activates the prompt.
• The font styling, background color, and lack of white-space all reduced the basic usability of the UI
• The text in the chat area lacked sufficient readability due to typography choices and contrast issues

Above: Excerpt from the Expert Review showing a Solution slide.

Collaborative Client Workshops

In addition to the Expert Review, a series of client workshops in Miro were conducted, led by our software architect and myself. We helped our client refine their product backlog, quickly ideated possible features, and provided overall guidance to the client team regarding their product roadmap. Our suggestions were being implemented almost in realtime by the client’s dev team.

Impact

This project was done quickly and showed what can happen when you get everyone on the same page. The client was able to utilize our feedback to make the best possible chatbot for their employees. Feedback from employees after implementation was positive and usage of the chatbot significantly increased due to the changes made.