top of page
Search

User Research & Prototyping - AI-Service Deletely

On average, people have around 168 online accounts, many of which are inactive. Each unused account increases the risk of data breaches and cyber-attacks. Yet, deleting accounts is often so complex and inconsistent across platforms that most people never do it.

For a course in Research Methods for AI-based Service-Design, I explored how AI could assist users in organizing and managing their online accounts more effectively.


Research Focus

I set out to understand:

❓How do people currently keep track of their online accounts?

❓When and why do they decide to delete accounts?

❓What challenges do they face in the process?

❓How do emotions shape their experience of digital account management?


Methods

To answer these questions, I combined two approaches:

📝 Quantitative survey with 15 participants (aged 21–60) to capture attitudes and self-reported behaviors

👀 Think-aloud observation of a participant deleting accounts, to study frustrations and decision-making in real time


Key Insights

  • Most users vastly underestimate the number of accounts they have

  • Security and privacy are the primary motivators for action, but 33% never delete inactive accounts

  • The deletion process is often so confusing that users give up entirely

  • 80% manage accounts manually; very few use tools or automation

  • The experience evokes frustration, confusion, and resignation


Design Requirements

Based on these findings, I defined requirements for a solution:

  • Tracking and organizing accounts automatically

  • Notifications when accounts are inactive or at risk

  • Easy, standardized deletion options

  • Minimal cognitive effort for users

  • Strengthened cybersecurity


Inspirations

I analyzed existing AI assistants such as Cleo (finance), Aha (mental health), and ChatGPT to identify useful patterns:


  • Proactive reminders and personalization

  • Emotional, conversational feedback

  • Step-by-step guidance for concrete actions


AI Concept: Deletely

I proposed Deletely, an AI-supported service that tracks accounts, identifies inactivity or security issues, and communicates with users through a chat-based interface. I visually presented the concept through 4 use-case scenarios.


  • The system detects unusual inactivity or data breaches

  • It informs the user via a generative AI chatbot

  • Users can ask questions, obtain background information, and make informed decisions

  • The AI can then initiate deletion, update rules, or re-categorize the account

    • This balances high automation with high user control, ensuring both efficiency and trust




Ethical Considerations

  • Data security and privacy are critical, as deletion often involves sensitive information

  • The system must provide explainability and user control to avoid misuse or blind automation

  • Strategies like two-factor authentication and user-defined data sharing are essential


Reflection

This project highlighted how human-centered AI design can reduce digital clutter while supporting cybersecurity. By combining user research, design requirements, and ethical reflection, I created a concept for an AI service that empowers users to take control of their digital footprint.



Process Presentation

During the development of Deletely, I presented my work process to other class members. The goal was to test and share different AI services during the process to evaluate them. I chose to try making Storyboards for the use cases with ChatGPT, which sparked a discussion on the following topics:

  • Are LLMS capable of really creating useful storyboards?

  • Is there a bias, because all characters look more "male-ish)

  • Is it more time-effective to prompt AI or to just sketch storyboards by hand?


Check out what ChatGPT came up with here ⤵️


Comments

bottom of page