top of page

Luan Pinto

Macbook Air M2 Silver Flatten.png

UX Research and Automation - Driving Down Support Costs by 98%

My Role

Product Designer

Project Timeline

4 Months (Research, Validation, MVP and V1.0 Launch)

Key Tools

Teams (Interviews)

ChatGPT (Transcripts/Analysis)

Notion

Figma (Data Tabulation/UI)

Figma Slides (Presentation)

Key Outcome

98.3% drop in initial support tickets and a 97% increase in new user satisfaction rating.

Problem Statement

The legal process automation software was suffering from a critical business problem: a high volume of support tickets originating specifically from new users (less than 3 months on the platform).

The Impact: This led to high costs, risk of churn, and a massive drain on support and coordinator teams. The hypothesized cause was a lack of initial guidance, non-intuitive interfaces, and technical terminology.

The Possible Solution: Structure and conduct in-depth qualitative research to diagnose the root cause of user confusion and propose a high-impact, low-cost solution (Chatbot) to achieve immediate self-sufficiency.

Challenge and Task

The Business Challenge: To address the symptom (high ticket volume) by focusing on the root cause (new user anxiety/confusion). The challenge was to prove the value of a UX-led solution to leadership and secure support for a product vision (Chatbot) that was previously only an idea.

​

The Task:

  1. Lead Discovery: Structure and conduct qualitative research (15 interviews) with key users to diagnose the actual cause of platform anxiety.

  2. Validate Strategy: Convert qualitative data into compelling business metrics to prove the urgency of the solution to stakeholders.

  3. Design the Onboarding Solution: Design and implement the Chatbot MVP as a quick, low-cost path to user self-sufficiency.

Process and Actions

Step 1: In-Depth Discovery and Synthesis
 

  • Method: 15 semi-structured interviews with users (0-3 months of experience), organized via Notion and conducted via Teams.

  • Analysis: Interview transcripts were processed (using ChatGPT for speed) and tabulated in Figma by thematic pillars (Usability, Learning, Terminology).

  • Key Insight: The main complaint was feeling "lost and afraid" due to the software's complexity, leading to the demand for on-screen help and manuals.

 

Step 2: Pitching and V1 Implementation
 

  • Strategic Action: Presented findings focused on the ROI of self-service to leadership (CEO, P.O., Managers) using Figma Slides, securing buy-in for the solution.

  • Solution Strategy: Implemented the Chatbot V1, initially powered by an encyclopedia of instructional content (a quick, low-cost solution).

  • UI/UX Design: Designed a bot character that conveyed friendliness (for welcoming new users) and authority (for conveying knowledge), differentiating it visually from the platform UI.

 

Step 3: Deployment and Validation
 

  • Collaboration: Excellent collaboration with the Development team.

  • Validation: Error flows (fallback mechanisms) were designed, and the solution underwent intense user testing before launch.

 
Matriz CSD TIME UX.png
Results and Impact

  • Support Metric (Core Outcome): An immediate drop of 98.3% in support tickets for initial queries.

  • User Satisfaction: A significant increase of 97% in user satisfaction rating among new users.

  • Product Evolution: The success of the V1 Chatbot validated the investment, allowing it to evolve into a Generative AI Assistant that now aids legal operators with complex tasks (drafting emails, processing terms, creating petitions).

Learnings and Next Steps

Key Learnings: The research proved that the problem is rarely where the business thinks it is. UX transformed a costly operational symptom (high support volume) into a productivity lever by addressing the core issue: user anxiety. The strategy of using a validated MVP (manual content) to prove ROI and then scaling to Generative AI was the key to securing long-term strategic investment.

Next Steps:

  1. Proactive Integration: Evolve the AI Assistant from an on-demand tool to a proactive system that suggests legal terms or pre-fills data in real-time, based on the context of the screen.

  2. Content Optimization: Implement continuous monitoring of AI fallback data to constantly refine the knowledge base and ensure the system's learning cycle is perpetual.

 
bottom of page