Chatbot

Project Overview

Over the years, the feature has undergone multiple iterations with a constant goal of enhancing the user experience, streamlining processes, and ensuring maintainability. Extensive research and staying updated with technology trends guided the project's direction and prioritization. Continuous creation, improvement, user feedback, and intelligent design have been essential for the enduring success of this feature.

Deliveralbes

Project Requirements

End User Experience

The goal is to enhance the End Users' experience, ensuring less frustration and more accurate results, ultimately establishing greater trust in the system.

Data Science

Collaborating with Data Science to automate responses, reduce manual effort, and streamline workflows.

Road Mapping

While designing and making decisions on the next version of the chatbot, prioritize choices that are future-oriented, ensuring the feature remains adaptable to evolving needs.

Iterating

Bring an idea to life

Overseeing this project from ideation to version 1 and beyond has provided invaluable insights into the rapid evolution of technology. After its launch, questions like "why did we do that?" surfaced, leading me to delve into my memories and past research. It was then that I recognized how what was once a standard experience had swiftly become outdated and a hindrance to our users. This realization prompted the need for a dedicated team for the chatbot, enabling quicker decision-making and implementation of necessary changes.

Feedback

Listening to the users

As feature adoption grew, I saw firsthand how much users struggled with setting up, maintaining, and trusting the accuracy of the chatbot’s responses. To tackle this, I worked closely with our customer service and support teams, diving into feedback and feature requests to pinpoint pain points. While the machine learning model performed well, it still required too much manual intervention, adding extra burden on users. My team focused on a solution by partnering with Data Science and leveraging AI to improve response accuracy, automate interactions, and minimize user effort by utilizing available documents. Additionally, we took an underutilized feature and enhanced the UI to reduce development time and increase feature adoption—ultimately making the chatbot a truly reliable and efficient tool.

Reflections

Emphasizing user flows

Initially, user flows were perceived merely as tools for designers, seemingly irrelevant to the product team's decision-making process. As I focused on enhancing the End User experience, the necessity of involving the entire team in these discussions became evident. Reviewing the current flow, evaluating user steps, and decision-making frequency led to a visual understanding of user effort, guiding us toward a better experience. Despite considerable debate on the new flows, my design choice wasn't selected. Project handovers and launch preparations left us questioning our decisions. As the only consistent team member, I had to explain the reasoning behind specific choices. Scrutinizing the wireframes exposed our mistakes. This process deepened the product management team's understanding of the value of our user flows.

Lessons Learned

Chatbots have had a reputation of being either highly successful or unreliable failures. The introduction of ChatGPT has necessitated flexibility in our future plans, complicating the upcoming launch. However, through brainstorming and research, I believe a seamless approach is achievable. Our commitment to the chatbot's success, coupled with consistent small updates, has fostered a trusted partnership between our company and our clients.

Pivoting as technology evolves

Small frequent changes