BACKGROUND
Dynamic risk control and customized leverage in prediction markets.
Driven by the industry's shift toward artificial intelligence, Alpha-Insight addresses a critical pain point in prediction markets: information overload. In the fast-paced environment of prediction markets, users frequently struggle with fragmented, rapidly changing data that creates cognitive friction and paralyzes decision-making.
As the lead designer for the end-to-end experience, I utilized progressive disclosure to balance depth and accessibility for a highly diverse user base. The interaction begins with a contextual action button beneath the K-line chart, revealing a bottom sheet of digestible, AI-generated insights and probability metrics. For deeper analysis, users can seamlessly transition into a conversational chat governed by strict compliance guardrails. Ultimately, by translating complex AI capabilities into an intuitive, layered mobile interface, this design successfully decreased time-to-decision, efficiently resolved routine queries, and drove sustained daily engagement.
Industry
FinTech
Prediction Market
Sports Tech
Platform
Mobile
PC
Android
Main Tool
Figma
Claude
Lean UX
Deliverable
UX Research
Competitor analysis, conversational UX analysis, information architecture, task analysis, edge case mapping, user flow mapping.
UX Design
Progressive disclosure strategy, interaction design, high-fidelity prototyping, AI state management (generating, typing, error states), design system integration.
Tech Handoff
Design-to-engineering handoff, latency & loading state handling, compliance guardrail logic, user acceptance testing (UAT).
STRATEGIC ALIGNMENT
Unlock user intuition with flexible market mechanics.
In spearheading the Alpha-Insight feature, our primary focus revolved around mitigating information overload in prediction markets to achieve three overarching objectives.
To realize this vision, we aligned our efforts around three core pillars:
1
Minimize cognitive friction by delivering contextual, easily digestible market data.
2
Resolve routine market queries through conversational AI.
3
Establish competitive differentiation by covering broader forecasting scenarios.
Research sets the stage.
THE ACTION-ORIENTED APPROACH
From fragmented data to centralized clarity.
Through semi-structured interviews, we mapped the traditional, multi-step journey of gathering market data. This validated our core hypothesis: users were suffering from severe cognitive overload. It proved that a centralized, progressive AI assistant would drastically reduce the mental friction required to make a trade.
COMPETITIVE BENCHMARKING
Focus on the market gap.
I audited the broader fintech and Web3 landscapes to understand how competitors were integrating AI. This established baseline user expectations and highlighted clear opportunities to differentiate our interface from existing market solutions.
CONVERSATIONAL HEURISTICS
Design for system trust.
I conducted a targeted analysis of conversational interface best practices—focusing on state management, prompt affordances, and error recovery—to ensure the assistant felt instantly responsive and inherently trustworthy.
UX builds the frame...
INTERACTION PRINCIPLES
Establish the rules of engagement.
To ensure the AI felt like a seamless utility rather than an intrusive gimmick, I anchored the architecture in three principles:
Data-First Focus (prioritizing core market mechanics over AI novelty);
Absolute Clarity (delivering synthesized, actionable insights instantly);
Unobtrusive Assistance (keeping AI entry points visually quiet).
CONTEXTUAL DATA HIERARCHY
Architecting information for immediate impact.
To effectively mitigate cognitive overload, I structured the data based on decision-making urgency.
The AI-Simulated Implied Probability and the synthesized AI Summary were placed at the apex of the hierarchy. Secondary drivers, such as real-time pricing factors and historical context, were mapped below. This deliberate stratification ensures that users receive high-value, actionable foresight the moment the bottom sheet is triggered, without having to dig for it.
PROGRESSIVE INTERACTION FLOW
From passive view to active chat.
The transition from the glanceable bottom sheet to the dedicated chat interface was mapped to support a fundamental shift in user intent.
To bridge this gap and eliminate "blank canvas" paralysis, I architected a system of contextual prompt chips that guide the user directly from the summary view into the conversational state. This interaction logic ensures a fluid transition, allowing users to move from reviewing high-level probabilities to probing complex market scenarios without breaking their cognitive momentum.
... and prototype brings it to life.
CONTEXTUAL ACTIVATION
Bridging the gap between observation and insight.
To minimize cognitive friction, Alpha-Insight is invoked directly from the primary chart view. A quick-access floating button beneath the K-line triggers a bottom sheet overlay.
This non-destructive entry point delivers immediate, AI-synthesized market context without forcing the user to abandon their current analytical flow.
STRUCTURED EXPLORATION
Triangulating data through progressive disclosure.
Prediction markets are inherently noisy. To streamline the user's investigation, complex data streams are categorized into three distinct pillars: Drivers & News, Historical Context, and Position & Edge.
A persistent header anchors the core probability and AI summary, allowing users to seamlessly toggle between tabs and synthesize micro-metrics without losing sight of the macro narrative.
CONVERSATIONAL ESCALATION
From passive consumption to guided inquiry.
The transition from browsing to conversing is facilitated by contextual suggestion chips. These pre-populated prompts lower the barrier to interaction, guiding users toward high-value queries based on the data they are currently viewing.
Crucially, the chat interface integrates strict AI guardrails, ensuring that responses remain objective, data-backed, and safely within the bounds of financial compliance.
Test. Validate. Iterate.
USABILITY FRAMEWORK
Validating mental models through targeted qualitative sessions.
To evaluate comprehension and navigation prior to launch, I conducted moderated think-aloud sessions with a curated mix of retail traders and prediction market veterans.
The primary objective was to assess whether our tripartite data structure naturally aligned with their existing research workflows and decision-making processes.
ITERATIVE REFINEMENT
Reducing cognitive load in complex data visualizations.
The insight from early testing was clear: users needed immediate momentum indicators, not just numbers.
I iterated by reducing the visual weight of the charts and introducing a plain-text synthesis—such as "Massive sell wall detected"—to instantly anchor user comprehension before they analyze the underlying graph.
FORWARD-LOOKING METRICS
Establishing KPIs to measure post-launch UX efficacy.
While qualitative testing validated our core interaction model, long-term success requires rigorous quantitative tracking.
Upon release, I will monitor the engagement rate of the suggestion chips, cross-tab navigation frequency, and overall time-to-insight. These metrics will serve as our baseline to ensure the interface continuously reduces the friction of market research.
