Aurelius
An AI-powered journaling app showcasing a shift in software design from tools to relationships.
Summary
How does AI change the way we design digital products and experiences? How can we use AI to help users reach their goals faster than before?
„Aurelius“, an AI-powered journaling app, showcases a shift in software design: from tools to relationships. I explore product strategy for AI products, their evolution and value proposition, as well as patterns and examples to build trust and collaboration in AI products.
Intro
The rise of AI brings both challenges and opportunities to design. As designers, we have access to new tools that change how we work. As users, we experience products that use AI to deliver new, powerful experiences, changing our expectations of what software can do. This, in turn, changes how we will be expected to design products in the future.
I’ve experimented with AI in my design work and now use it regularly. But after working on a project last year where AI was central, I wanted to go deeper and design a product that truly uses AI to create new, transformative experiences.
This case-study is the result of that exploration. It aims to understand what makes a good AI product, how AI can improve products and businesses, and how user experience and product design will evolve in the coming years.
Disclaimer: The product presented here is fictional and exists only as concepts, designs, and prototypes. AI was not involved in this project beyond proofreading. This project is about designing products in the age of AI, not designing with AI.
The central shift: how AI changes software design
Traditional software is designed as a tool. Interfaces and systems are carefully planned to help users complete specific tasks as efficiently as possible.
As designers, we optimize screens and flows, reduce friction, refine copy, and improve completion rates. The assumption is simple: users know what they want, and software helps them do it faster.
But humans don’t think in isolated tasks. We think in projects and goals that change and evolve over time.
The rise of large language models promises to close this gap.
AI software understands natural language, keeps context, and anticipates needs. Instead of merely helping users execute tasks, it can suggest what to do, how to do it, or even do it on their behalf. It starts to feel less like a tool and more like a relationship.
With memory and behavioral learning, AI can treat users like humans with changing goals. Dynamic interfaces reinforce this understanding, while initiative and judgment signal care. Together, these qualities help build trust, an increasingly important differentiator.
Thanks to AI, software can finally learn and become more useful over time. But designing relationships instead of tools requires a new mindset.
Design challenges in the age of AI
When designing products with memory, initiative, and judgement, a different design approach is necessary. The core principles of user experience design stay relevant: understanding users, mapping goals and intent, designing flows. But instead of designing a static interface and fixed system, we need to design a collaborator — dynamic, evolving, and adaptive.
A key challenge is trust. With traditional software, trust came from reliability, familiarity, and brand. With AI, trust develops more like a human relationship.
I remember my first time using ChatGPT. I was skeptical, expecting another shallow chatbot. But with each successful interaction, my trust grew. As reliability increased, so did my willingness to delegate authority.
AI products must strike a careful balance: powerful enough to be genuinely helpful, yet transparent enough that users feel in control, even as the system acts autonomously.
Designers must focus less on screens and more on relationships: building trust, communicating confidence, and creating interfaces that adapt to individual users.
The state of journaling
I’ve kept a daily journal for ten years. It’s one of the best habits I’ve ever formed, helping me reflect, build empathy, and wind down. Sometimes, I write freely. Other times, I respond to prompts. For me, journaling is a tool for mental health and personal growth.
I use a notebook and a pen to avoid distractions, but the analog approach has drawbacks: physical storage, manual search, and the need to always have the materials on hand. Discovering patterns across years of entries is difficult.
I’ve considered digital alternatives. Searchable entries and prompts are appealing, yet none have convinced me to abandon my notebooks. This led me to explore what a digital journaling product could look like in the age of AI and what unique value it could offer.
Problems and opportunities
An AI-native journaling app (called „Aurelius“) could address several challenges:
- Keeping journaling fresh and engaging over time.
- Providing guidance tailored to personal goals and situations.
- Surfacing insights hidden across years of entries.
By analyzing journal entries, Aurelius could identify patterns and generate insights that grow more valuable as it learns about the user.
The evolution of AI experiences
Starting from an analog journal, I outlined how a digital journaling app could evolve into an AI-first experience. This evolution is driven by context.
As products better understand who their users are and what they want, they can evolve from tools to assistants, collaborators, and partners.
Level 1: A digital journaling app
At its simplest, a digital journal replaces pen and paper. Entries are searchable, storage is effortless, and reminders or guidance support habit formation.
However, the initiative still lies with the user. Personalization is limited, and the product remains a tool.
Level 2: A conversational journaling app
Adding AI enables natural language interaction. You can talk to your journal, ask questions, and retrieve past experiences with ease.
Yet interactions still start from scratch, and the user remains responsible for directing the experience.
Level 3: A task-aware journaling app
Here, the app begins to act when asked. It can suggest prompts, analyze entries, and proactively support goals.
At this stage, trust becomes critical. The system must demonstrate understanding and usefulness when taking initiative.
Level 4: A personally intelligent journaling app
Now the initiative shifts to the system. Aurelius remembers preferences, learns rhythms, and adapts over time.
It nudges, helps, and adjusts without being asked, and it feels respectful and transparent. The product no longer feels like a tool, but like a relationship.
Patterns to transform Aurelius into an AI-first experience
Studying state-of-the-art AI products reveals patterns that help build trust and collaboration.
Clarify intent
AI can appear overconfident. Clarifying intent before acting shows understanding and prevents misalignment.
Aurelius summarizes plans, offers feedback, and asks for approval before making changes, thus mirroring healthy collaboration.
Make reasoning visible
Trust depends on understanding why decisions are made. Exposing logic, sources, and uncertainty builds confidence.
Aurelius explains its suggestions and allows users to explore the underlying reasoning.
Make users feel safe
Autonomous systems must be reversible. Clear undo and override options encourage exploration and trust.
Aurelius always provides ways to revert changes it makes on the user’s behalf.
Invite users to work together
Collaboration feels better than delegation. Asking for feedback and refinement keeps users engaged and in control.
Aurelius treats planning as a shared process and invites users to work together.
Have a transparent, editable memory
Personalization requires memory, but memory must be visible and controllable.
Users can view, edit, delete, or fully reset what Aurelius remembers.
Adapt the interface
Fixed interfaces can limit AI. As goals and contexts change, the interface should adapt.
Aurelius steps back during focused writing and becomes conversational when guidance or reflection is needed, adjusting to the moment. It shows relevant context directly during conversations.
The business case for relationships
As AI becomes commoditized and natural language becomes the interface, differentiation shifts.
The true moat will be the relationship: trust, time, and accumulated context.
Features can be copied. Interfaces can change or disappear. But a deep understanding of the customer and trusted behavior cannot. Users will choose products that act in their best interest—products they trust. And trust, once earned, compounds. In the future, a trusted relationship will be the decisive advantage of your product.
Conclusion
While the future of AI remains uncertain, large language models and natural language interfaces are here to stay. Product design must evolve accordingly.
Designing for trust, relationships, and collaboration represents a major opportunity for designers and a necessity for businesses that want to remain relevant.
Further reading
In case you want to learn more about design patterns for AI experiences and how to build trust, check out these sources: