Project Overview
This case was based on a brief from a company developing digital tools for the real estate industry. Their platform uses a digital representation of buildings, floors and spaces to help people navigate and manage physical workplaces that they call a Digital twin.
The challenge was to explore how this technology could make everyday work easier in a flexible, activity-based office. I was given four potential user needs and asked to choose one to explore further:
Helping employees find a suitable workspace
Locating meeting rooms
Seeing which colleagues are working on-site
Report something broken
With limited information about users and context, I was asked to define a direction, explore alternative solutions and develop a concept within a short timeframe, instructions said it wouldnt take more than half a day.
I used the case to explore how AI could accelerate this process while of course keeping the key design decisions human-led.
I owned the process from interpreting the brief and defining the problem to exploring concepts and developing the final prototype.
Problem framing
UX & interaction design
Concept development
Prototyping
AI-assisted design exploration
0️⃣
Outcome
A high-fidelity concept exploring a smarter way to find and choose a workspace in an activity-based office.
The final prototype demonstrates how a digital representation of the workplace could guide employees towards a suitable workspace based on their needs, rather than simply showing available spaces.
The project became a turning point in how I think about AI in my design process, not as a replacement for design thinking, but as a way to explore and iterate at a much higher pace.
1️⃣
Core Challenges
Working with limited information
The brief intentionally left gaps around users, context and behaviour. I had to make assumptions while identifying what I would want to validate through research.
Choosing where to focus
There were four potential user needs. With limited time, I had to decide which problem would create the most interesting opportunity to explore.
Using AI without outsourcing the thinking
AI could generate possibilities quickly, but the challenge was knowing what to keep, what to reject, and how to turn promising ideas into a unified product experience.
2️⃣
Design Process
Frame
The brief presented four potential user needs, but only enough time to explore one of them in depth. I therefore created a simple scoring framework to evaluate each opportunity against three criteria. I scored each criterion from 1–5 to make the trade-offs explicit and decide where to focus.
User value: How often and how much the need affects the employee’s workday.
UX potential: How many different user needs and challenges there are to explore.
Connection to the Digital twin: How much value can be created by using information about the building.
This gave me a clear direction for the rest of the case and allowed me to focus the limited time on the opportunity with the strongest combination of user value, UX potential and connection to the Digital twin.
Explore
With “Find an available desk” selected as the direction, I explored the problem space further by looking at the different situations, behaviours and needs involved in finding a suitable workspace.
I used AI to challenge my initial assumptions, explore different scenarios and look at the problem from multiple perspectives. This helped me uncover additional user needs and potential problems that I might not have identified on my own, opening up the problem space and giving me more areas to explore.
After a 15-minute discussion with AI, I asked it to summarize the key points we had explored and turn them into four clear outcomes.
Explore yourself with maximum control: See the office/map and choose a workspace yourself.
B. Describe your needs and shared control: "I’m running a workshop for two hours” →
the system finds a suitable workspace.
C. Choose for me with minimal effort: "I want to work undisturbed” → the system
recommends a workspace.
D. Minumum effort: "I don’t mind, just find a good workspace for me.”
The system uses previous behaviour and actual information.
This gave me a structured starting point for the next phase to help me identify which areas were worth exploring further.
With four potential directions identified, I created a matrix and evaluated them from two perspectives: "user value vs. user effort" and "user value vs. implementation effort".
The first matrix helped me understand which directions could create the most value for users relative to the effort required from them. The second helped me consider the potential implementation effort and challenge some of my assumptions about what might be technically feasible.
Since I didn’t have access to technical data or the opportunity to discuss the ideas with a developer, the implementation effort was based on assumptions rather than validated constraints. From a UX perspective, I therefore chose to prioritize the direction with the highest potential user value and lowest user effort.
This meant deliberately advocating for the user while being transparent about what would still need to be validated from a technical and product perspective.
The final decision landen in D. Minimum effort.
Ideate and Prioritize
With the problem direction selected, I turned it into a How Might We question:
How might we make it easy for employees to find the right workspace without having to spend unnecessary time and energy choosing?
I set a 15–20-minute timer and used the question to generate as many ideas as possible without judging them too early.
Once I had a range of potential solutions, I used a simple dot-voting exercise to prioritize the ideas with the strongest potential. Three directions stood out and became the starting points for the next stage of exploration.
Sketch
With three directions selected, I started sketching by hand. I used the Crazy 8 method to visualize and expand the ideas quickly, turning the concepts from text into tangible interface possibilities. The goal wasn’t to create polished solutions, but to explore different ways the ideas could work and uncover new possibilities through sketching.
Refine
I then took some of the strongest ideas into a conversation with ChatGPT. Rather than asking AI to design the solution, I used it as a sparring partner: discussing the concepts, challenging different approaches and exploring how they could evolve. Once the direction became clearer, I asked AI to turn our discussion into a structured design prompt that captured the key decisions and interaction ideas.
Make tangible
I used the resulting design prompt as a starting point in Figma Make, turning the concept into a tangible interface that I could interact with and iterate on.
This created a fast loop between thinking, sketching, discussing and building. Allowing me to move from an early idea to something concrete that I could evaluate and refine.
Sketch
Refine
Make tangible
3️⃣
Visual design, UI & Delivery
The exploration led to a concept focused on making it easier for employees to find the right workspace with minimal effort.
The final prototype brings the chosen direction to life, showing how the workplace could be represented visually and how information about spaces, availability and employee needs could help guide the user towards a suitable workspace.
Rather than presenting a finished product, the prototype is a tangible expression of the concept — designed to make the idea easier to experience, discuss and evaluate.
Key benefits and design decisions included:
Reduced cognitive load – users don’t have to search for an available workspace themselves.
Personalized recommendations – recommendations adapt to the user’s needs and context for the day.
Faster decision-making – users get a relevant suggestion immediately instead of browsing multiple options.
Confidence before booking – users get enough context to understand and trust the recommendation.
Seamless experience – a connected flow from recommendation → booking → navigation.
4️⃣
AI as a design partner
🧑💻
VICTOR
Direction
Decisions
Prioritazion
🤖
AI
Challenge
Expore
Generate
The goal wasn’t to let AI design the product. It was to use AI to increase the breadth and speed of my exploration while keeping the design direction and decisions human-led.
AI helped me
Challenge assumptions — question my initial thinking and expose alternative perspectives
Broaden the solution space — generate directions I might not have considered
Explore scenarios — quickly work through different user situations and possibilities
Structure ideas — turn scattered thoughts into clearer concepts and alternatives
I stayed responsible for
Choosing the direction
Prioritizing ideas
Making design decisions
5️⃣
Interviews and user testing
Do employees actually want the system to choose a workspace for them, or would they prefer to make the final decision themselves? Interviews and user testing would help uncover their expectations, preferences and level of control.
Observation and interviews
Which workspace characteristics matter most when employees choose where to work? Observing how people use the office, combined with interviews, would help identify the factors that influence their decisions.
Technical discovery
What data is currently available, and what additional sensors or integrations would be needed to support workspace recommendations? Exploring the existing digital twin with Pythagoras would help establish the technical possibilities and limitations.
Scenario testing
How should the system respond when people change plans, leave a workspace or no longer need a reservation? Testing different scenarios would help define rules for reserving, releasing and updating workspaces.
Prototype testing
Do users understand why a workspace is recommended, and do they trust the recommendation enough to act on it? Prototype testing would help identify usability issues and reveal where the experience needs to be clearer.
6️⃣
Reflection
This project showed me how AI can accelerate the design process by helping me explore more directions, challenge assumptions and turn ideas into tangible concepts faster.
The key takeaway: AI can expand the possibilities, but the direction, priorities and final decisions remain mine.