AI ChatBot

An AI-powered assistant designed to streamline developer support, reducing inquiry resolution time and improving workflow efficiency.

MY Role

UX/UI Designer

Team

Front-end Developers

Product Manager

Stakeholders

Tools

Figma

summary

In modern software development, quick access to accurate information is critical. Developers often struggle to locate key resources within extensive documentation, leading to inefficiencies, miscommunication, and delays in resolving technical inquiries.

To address this challenge, I led the design of an AI-powered chatbot aimed at streamlining developer support, minimizing search time, and fostering seamless collaboration.

The Impact

The AI chatbot boosted developer efficiency by increasing task completion rates from 8-9 to 12-13 per hour, slashing inquiry resolution time from over 30 minutes to about 1 minute, and transforming the experience from frustration to productivity and satisfaction.

The Old vs The New

Developers struggled with delays caused by the contact page, which often took too long to provide replies or resolve issues. Feedback and analysis highlighted the need for an instant, self-service solution—leading to the development of an AI-driven support strategy.

Below is a side-by-side comparison of the old support system vs. the AI chatbot, showing the difference in response times and user experience.

What Really Matters?

I defined key metrics to measure existing pain points and evaluate the impact of our solution upon completion.

# of tasks
completed per hour

8-9

Inquiry resolution time

> 30 minutes

Common Feeling

Frustrated
and Inefficient

Strategizing the AI Solution

I selected ChatGPT for its contextual accuracy, scalability, and seamless integration with the existing developer portal. The chatbot needed to handle complex queries efficiently while maintaining brand consistency.

Enhancing ChatBot Intelligence

To improve usability, I implemented context retention, enabling the chatbot to remember previous questions within a session and offer personalized recommendations based on prior interactions. For example, if a developer repeatedly asked about certain code snippets, the chatbot would proactively surface related FAQs and suggest best practices before the user even finished typing their question.

Designing A ChatBot That Stands Out

I mapped user conversations to streamline interactions, ensuring clarity and error handling. The chatbot UI featured microinteractions like typing indicators, a sleek glassmorphism design, clear AI/user differentiation, and WCAG-compliant accessibility for inclusivity.

Testing It Out

I created Figma prototypes to test chatbot interactions and conducted A/B testing to refine accuracy and usability. Developer feedback guided iterative improvements for an optimal experience.

The Results

Beyond answering queries, the chatbot streamlined developer workflows by integrating with GM’s API Library. Developers could retrieve code snippets, auto-fill API request parameters, and debug errors directly within the chat interface. This eliminated the need to search through multiple documentation pages, reducing lookup time.