Show Me What Changes
Explored a different role for AI in financial decisions: not choosing for the user, but making the consequences and trade-offs of each choice easier to understand. Designed a contextual decision-support experience for BJAK through rapid research, concept development and mobile prototyping.
Project Overview
BJAK · Product Design Challenge
AI Decision Support for Financial Choices
A 3-day independent concept exploring how contextual AI could help users move from comparing financial products to understanding what those differences mean for their own situation.
Responsibility: Problem framing, rapid user research, product concept development, UX flow & interaction design, wireframing, UI design, prototyping, validation planning
Keywords: Product Design · UX Research · AI Interaction · Decision Support · Information Architecture · Mobile UX · Prototyping
the challenge
''Comparison gives users information.
But does it help them understand what the choice means for them?''
BJAK already helps users compare financial products and understand differences in price, coverage and limits.
But a difference in coverage does not automatically tell a user why that difference matters in their own situation.
The challenge was to explore whether AI could bridge this gap by helping users connect product differences with real-world consequences without taking the decision away from them.
DISCOVERY
Before designing the AI, I challenged the problem.
My initial assumption was that users needed help choosing between financial products. A lightweight user check was used to test whether the real difficulty was choosing or understanding what the differences meant.
Research question
Do people struggle to choose or to understand what their choices mean?
Method
n = 5–7 participants
10–15 min / participant
Exploratory interviews + decision task
Decision task
Compare two travel insurance plans and explain which one you would choose and why.
Context shift
Now imagine you're travelling with an SEK 13,000 camera.
Would your choice change? Why?
THE CONCEPT
From comparison to consequence.
The research shifted the problem from “How might we help users choose?” to “How might we help users understand what changes when they choose?”
Instead of replacing comparison with an AI recommendation, I introduced a decision-support layer that connects product differences to the user's situation.
The decision-support model
An AI-powered decision-support experience that connects a user's context with verified product information to make the consequences and trade-offs of different financial choices easier to understand.
Design principles
AI supports the decision. The user owns it.
-
Explain, don't persuade. AI makes trade-offs visible rather than pushing users toward a “best” option.
-
Personalize meaning, not facts. Product information remains grounded and transparent; AI adapts how its relevance is explained.
-
Keep the user in control. AI supports understanding and exploration. The final decision remains with the user.
THE EXPERIENCE
AI as a decision-support layer, not the decision-maker.
The experience follows a focused decision journey: first establish context, then compare relevant options, surface the difference that matters, and finally make the trade-off concrete.
The Journey
01 — CONTEXT
Capture only the information that can change how the comparison is interpreted.
↓
02 — COMPARE
Keep the underlying product comparison transparent.
↓
03 — AI INSIGHT
Surface the difference that is most relevant to the user's situation.
↓
04 — WHAT-IF
Turn an abstract difference into a concrete scenario.
↓
05 — DECIDE
Make the trade-off explicit and return the decision to the user.
The AI layer sits between comparison and decision. It interprets verified product information through the user's context rather than replacing the comparison itself.
FINAL DESIGN
The final experience turns a traditional comparison into a sequence of increasingly contextual questions: What matters for this trip? What changes between the options? What does that difference mean? And finally, which trade-off am I willing to make?
AI doesn't tell the user which plan is best. It helps them understand what changes when they choose one.
01 — CONTEXT
Tell us what matters
Capture the minimum context needed to make the comparison more relevant.
02 — COMPARE
Plans that fit your journey
Keep price, coverage and limits visible while highlighting the differences that may matter.
03 — AI INSIGHT
Two things worth knowing
Connect the user's context with verified product information and surface the consequences worth understanding.
03 — AI INSIGHT
Two things worth knowing
Connect the user's context with verified product information and surface the consequences worth understanding.
04 — WHAT-IF
What changes if your trip is cancelled?
Translate an abstract coverage difference into a concrete scenario and make the trade-off visible.
05 — DECIDE
Your decision
Summarize why the selected plan fits the situation while making it clear that the final decision remains with the user.
THE KEY DESIGN DECISION
Make the trade-off visible — not the recommendation.
A conventional AI assistant might answer: “Plan B is better for you.”
I deliberately avoided that interaction. Instead, the system explains why a difference matters, lets the user explore a what-if scenario, and makes the cost of that choice explicit.
The goal is not to create an AI that decides. It is to reduce the cognitive work required to make a decision the user can understand and own.
VALIDATION & REFLECTION
The next question is not whether users like the AI. It is whether it improves decision understanding.
Because this was a 3-day concept exercise, the next step would be lightweight prototype validation focused on decision comprehension rather than feature preference.
Success signal: Users can explain their decision in their own words by using their context and the relevant trade-off, rather than simply repeating the AI's recommendation.
What I would change: If users understand the insight but question why the AI surfaced it, I would make the connection between user context, product data and AI reasoning more explicit. If users make the decision equally well without the AI layer, I would reduce or remove unnecessary AI interactions. If the AI creates additional cognitive effort, I would simplify the interaction rather than adding more explanation.
Reflection: The most important shift was moving from “help users choose” to “help users understand the consequences of choosing.” In a financial context, AI becomes more useful when it reduces interpretation effort without taking ownership of the decision.