Context
6 months project in 2021
Project type
Final major project for my MA
Role
UX Research
Team
Maria Seves and Eda Erdal
The decision-making tool for AI Personality
💭 The Goal
To answer a pressing question: how can we bring more transparency to human-AI interactions? Our work sought to improve how brands and designers craft AI personas, emphasising clarity, user engagement, and ethical design.
🪞 The Problem: AI Without Personality
An AI’s personality is more than its voice or dialogue structure—it represents the brand and directly impacts user trust and experience. Yet, the design process for AI personalities often lacks a structured framework. We aimed to bridge this gap with Backstory.
🎭 The Solution: An Interactive Design Toolkit
Backstory guides users through creating AI personas step-by-step, reflecting on:
- Goals: What purpose does the AI serve?
- Anatomy: How should it look and feel to the user?
- Personality: What tone and behaviours best represent the brand?
- Social Impact: What implications does this design have on users and society?
The process culminates in generating a customised AI Manifesto, a tangible artefact to align teams and ensure design consistency.
🔑 Key Insights
Starting this project from a single research question, we had the opportunity to explore different generative research methods in addition to evaluative ones.
Research & Methods
The kick-start for this project was a Research Question
How can we bring more transparency to Human-AI Interactions?
To answer this question, Eda and I partnered to investigate through design to understand some key questions: what does AI mean? How does AI make people feel? And why does Alexa look like a Magic 8 ball?
As all good UX goes, the process took us a couple of months, multiple UX Methods, a few coffee cups, and many iterations to reach our final design. Here is a glimpse of our design process.
Our research process was divided into two main categories: generative methods to explore possibilities and evaluative methods to refine our solutions.
Generative Methods
These methods focused on uncovering user needs, emotions, and creative possibilities for AI design:
- Concept Mapping: we explored AI's abstract nature, drawing comparisons between chatbots and fortune tellers to make the design challenges more tangible.
- Emotion Mapping Workshops: participants reflected on their emotional responses to AI, revealing excitement about its potential but concerns about privacy and ethical implications.
- Creative Workshops ("Make Your Own Alexa"): using Play-Doh, participants reimagined Alexa with features they valued most, such as emotional expressiveness, improved personalization, and accent adaptability.
- Crazy 8 Brainstorming and Storyboards: rapid ideation exercises produced diverse ideas, culminating in storyboards that visualized potential solutions.
Evaluative Methods
These methods validated and improved our designs through user feedback and iterative testing:
- Artefact Analysis: by studying Amazon Alexa’s design, we identified its strengths (futuristic aesthetic) and weaknesses (accent recognition and personalization gaps).
- Wireframing and Prototyping: iterative design—from paper sketches to high-fidelity prototypes—allowed us to refine the toolkit based on user insights.
- User Testing: 6 participants tested the prototype, providing valuable feedback on the structured journey and AI Manifesto. Testers appreciated its practical application for team collaboration and planning.
Key Insights
With all the research methods and different iterations, we generated over 10 insights - here is a shortlist of the top 5
- Transparency builds trust: users value tools that clarify AI’s functionality and personality, helping foster trust and understanding in human-AI interactions.
- Personalization: customization options, such as voice adaptability and accent recognition, are key to improving user satisfaction with AI tools.
- The role of personality: designing clear and intentional AI personalities, including tone of voice, enhances engagement and aligns with user expectations.
- Balancing anthropomorphism: While users appreciate emotional expression in AI design, overly human-like features raise ethical concerns and require careful consideration.
- Practical tools add value: the AI Manifesto generated through Backstory was praised as a tangible outcome that supports team collaboration, documentation, and planning in AI design processes.
Reflecting on our work 💭
Developing Backstory was an intensive, rewarding process. By helping designers make intentional choices about AI personalities, we took a step toward more transparent and meaningful human-AI interactions.
Try out the prototype ✨
Click on the Figma file to open the prototype and navigate our toolkit.
