Sales professionals, business development executives, professionals involved in lead generation and cold outreach.
UX research
UX design
UI design
Collaboration: Slack
Research: Google Analytics, Maze, Google Forms, Microsoft Clarity
Prototyping: UIzard, Figma, Tailwind
Luna AI had a major issue with a 52% churn rate, showing that most users did not continue using the platform after signing up and trying it once. This high churn rate was a big problem for the platform's success and long-term growth.
Solution included organising usability tests with five target audience users. This research-based approach helped identify critical areas for improvement, driving an iterative design process to refine the platform.
Implementing the redesign recommendations resulted in a 39% churn rate reduction, as evidenced by A/B testing. This improvement led to increased user retention, expansion of Luna AI's client base.
Lack of an onboarding process in LunaAI makes it difficult for new users to explore its features and functionalities, hindering their ability to effectively utilize the app.
Introducing an interactive guided tour during the onboarding process will enhance user understanding of the app's features and functionalities, leading to improved navigation and increased engagement.
An interactive guided tour that introduces new users to the app's key features and functions, providing step-by-step instructions and visual aids to facilitate a smoother onboarding experience.
Unclear information architecture within the app causes user confusion regarding various aspects of its functionality, resulting in a less intuitive user experience.
Implementing a more intuitive and streamlined information architecture, based on insights from card sorting tests, will lead to reduced user confusion, increased user satisfaction and retention.
Redesigned the app's layout based on insights from card sorting tests to create a more intuitive and user-friendly information architecture that facilitates effortless navigation.
Users experience frustration and reduced task efficiency when they have to search for necessary information within a large knowledge base during task completion. This negatively impacts their overall satisfaction and the effectiveness of their task execution.
Providing a more efficient way to access necessary information during task completion will lead to reduced user frustration and increased satisfaction.
Adding AI-powered user assistant that can give needed information in form of support chat and take actions on the user behalf.
The objective of the research was to gain insights into the experiences of new Luna users and identify areas for improvement. Five participants matching the target audience were recruited for usability testing.
Participants were asked to perform the following tasks:
After completing the usabiility testing, participants answered several follow-up questions to assess their experience more thoroughly. I compiled their responses into an affinity diagram and grouped them by common themes and insights.
To communicate my findings too the team, I've created a customer journey map
The findings from the usability interviews revealed that users encountered difficulties navigating the dashboard. To address these issues and refine the grouping, an open card sorting test was conducted using Optimal Workshop with 20 participants possessing a sales background but no prior experience with Luna.
Upon conducting a comprehensive research, I pinpointed several key vectors that could effectively contribute to reducing the churn rate. These areas include:
1. Creating a Structured Onboarding
Identifying the absence of a clear onboarding process in LunaAI, I emphasized the need for an engaging and informative experience to ensure users grasp core functionalities and recognize the application's value.
I adopted a benefits-oriented approach, focusing on the primary user flow—sending their first cold email. To achieve this, I incorporated various design patterns:
2. Refining the information architecture
Guided by findings from user research and heatmap analysis, I focused on refining the layout and labels for enhanced clarity and organization. Multiple areas demanded improvement, such as the unclear distinction between invited team members and email addresses for cold outreach, unclear categorization in settings, and the need for a more coherent visual hierarchy on various pages. To address these issues, I designed separate flows for managing team members and playbook-assigned accounts, reorganized settings categories for improved clarity, and optimized the visual hierarchy on key pages using heatmap data.
Based on the insights gleaned from the research, I created low-fidelity wireframes. I then organized meetings with my team to present the proposed improvements, explain the rationale behind them, and invite my colleagues to share their thoughts and feedback.
3. Making help easily accessible and using the user's language
Making help easily accessible and using the user's language: Analyzing the current state of the dashboard revealed that the "help and support" heuristic was not adequately addressed, as the knowledge base was presented as a list of hard-to-navigate articles. Also, it was unclear how users could obtain support—during usability testing 74% of users responded with "I am not sure" when asked how to seek assistance.
Ideally, we would utilize an AI model trained on the LunaAI knowledge base, capable of processing questions in natural language. However, due to practical constraints, we opted for CommandLine AI, an AI-powered user assistance tool that monitors user actions, provides support when needed, and adheres to privacy regulations like GDPR and HIPAA.
One direction LunaAI intends to pursue in the future is to focus on dialogue-based interactions between the user and the system. Currently, this feature is available on the platform as a help option for users unfamiliar with the interface. However, usability testing has shown that dialogue-based exchanges significantly enhance user satisfaction and task completion rates. Therefore, one of the primary tasks for the future will be to research and prototype potential solutions in this area
Working on this project sparked my interest in dialogue-based human-computer interactions. To learn more about this topic, I reached out to Inge De Bleecker, a senior Conversational AI and UX Consultant, through the ADPlist platform. I received a number of resources that she found useful and that could help me build a strong foundation in this direction. Currently, I am reading 'Dialogue Explanations for Rule-Based AI Systems' by Y. Xu. After that, I plan to take a self-paced course at the Conversation Design Institute.
AI-driven platform that automates the extraction of critical information from medical records for medicolegal cases.