Min Thu Kyaw.

Confidential Enterprise AI Product — Trading Intelligence and Decision Support

AI Trading Assistant for Blockchain Mobile App

Led the product strategy and high-level solution design for an AI-powered trading assistant that helps cryptocurrency users understand market conditions, analyze assets, monitor portfolios, generate research insights, and make more informed decisions while keeping all trading actions under user control.

Role
AI Product Manager & Solution Architect
Project type
Confidential Enterprise AI Product — Trading Intelligence and Decision Support
Status
Completed
Organization
Confidential Blockchain Company
Timeline
2026-02-01 — 2026-06-30
Team
Cross-functional collaboration with mobile engineers, backend engineers, AI and machine-learning engineers, market-data specialists, quantitative analysts, UX designers, security teams, compliance stakeholders, quality-assurance engineers, product operations, and business leadership. Team size, internal structure, and individual identities are withheld due to confidentiality.
Confidentiality
High — NDA-Protected / Sanitiz
Published
published

Confidentiality note

This case study is a sanitized and generalized reconstruction of a confidential enterprise AI project developed for a blockchain mobile application. It was recreated from memory solely to demonstrate product-management, conversational-design, AI-strategy, UX, and solution-architecture capabilities. It does not contain the company’s original product documents, source code, production prompts, customer records, portfolio data, transaction data, exact architecture, market-data contracts, internal APIs, model configurations, proprietary trading logic, security controls, vendor details, business metrics, or confidential operational information. All company references, user identities, interfaces, workflows, examples, market values, technical details, and product artifacts have been anonymized, simplified, simulated, or reconstructed for portfolio use. The product is presented as a market-research and decision-support tool, not as guaranteed financial advice or an autonomous trading system.

Project overview

This confidential enterprise project focused on designing an AI-powered market intelligence and trading assistance experience within a blockchain mobile application.

I worked across AI product management and solution architecture, translating trader needs into product requirements, conversational workflows, AI capabilities, market-data integrations, safety controls, and a high-level technical architecture.

The assistant was designed to help users explore market trends, understand technical indicators, monitor watchlists and portfolios, compare assets, receive personalized insights, and prepare potential trading plans. It combined conversational AI, retrieval-augmented generation, real-time market data, technical analytics, market sentiment, portfolio context, and configurable alerts.

The product was positioned as a research and decision-support tool. It did not guarantee returns, provide unquestionable financial advice, or execute trades without explicit user confirmation.

Due to confidentiality obligations, all portfolio materials are sanitized and reconstructed from memory. They do not represent the company’s exact production architecture, original documentation, prompts, models, trading logic, APIs, customer data, or proprietary implementation.

Business problem

Cryptocurrency traders must interpret rapidly changing information from prices, technical indicators, market news, sentiment, portfolio activity, and multiple trading pairs. This information is often fragmented across different applications and presented using terminology that can be difficult for less experienced users to understand. Traders may spend significant time collecting information while still making decisions based on incomplete context, emotional reactions, or isolated signals. The company needed a secure and explainable AI experience that could: Consolidate relevant market information Answer natural-language market questions Explain technical indicators in understandable language Personalize insights using watchlists and portfolio context Help users compare assets and trading scenarios Generate alerts for meaningful market changes Clearly communicate uncertainty and risk Keep final investment and trading decisions under user control

Product objective

Design an AI-powered trading assistant that helps cryptocurrency users research markets, understand asset behavior, evaluate potential opportunities, monitor portfolio exposure, and make more informed decisions through personalized and explainable insights. The product should improve access to market intelligence without presenting uncertain predictions as facts, guaranteeing financial outcomes, or removing user control from consequential trading actions.

Product approach

I used a user-centered, evidence-grounded, and human-controlled product approach. The product design began with the trader journey: opening the assistant, asking a market question, selecting an asset, retrieving relevant market and portfolio context, reviewing AI-generated analysis, exploring supporting indicators, preparing a possible plan, configuring an alert, and deciding independently whether to act. The assistant was designed as a layered research system: The user asks a market, asset, portfolio, or trading-related question. The system detects the user’s intent and extracts relevant entities. Authorized user context such as preferences, watchlists, portfolio holdings, and recent activity is retrieved. Current market data, technical indicators, sentiment, and approved news sources are collected. The AI analyzes the available evidence and generates a structured response. The response includes supporting factors, timestamps, uncertainty, and appropriate risk disclosures. The user may compare scenarios, create an alert, save research, or prepare a potential trading plan. Any consequential action requires explicit user review and confirmation. User feedback and outcomes support continuous product improvement. The design intentionally treated AI as a market-research and decision-support capability rather than an autonomous financial decision-maker.

Solution summary

The solution was designed as an AI-powered trading intelligence capability embedded within a blockchain mobile application. A secure conversational gateway manages authentication, sessions, rate limits, language handling, and communication with the mobile client. The orchestration layer handles: Intent detection Asset and market entity extraction User-context retrieval Workflow routing Tool and data-source selection Response generation Confidence and safety evaluation The market-intelligence layer can combine: Real-time and historical price data Trading volume and liquidity information Technical indicators Market sentiment Approved news and research content Watchlists and price alerts Authorized portfolio information User preferences and risk settings A retrieval and analytical pipeline provides relevant evidence to the language model. The model then produces structured market summaries, indicator explanations, scenario analysis, portfolio observations, and research guidance. The user can review supporting evidence, compare assets, save insights, configure alerts, or prepare a potential trading plan. Any trading action remains separate from the analytical response and requires deliberate user confirmation. Monitoring and governance components track data freshness, model quality, unsupported claims, user feedback, system performance, and safety incidents.

Target users

  • Retail cryptocurrency traders
  • Intermediate and experienced digital-asset investors
  • Users monitoring cryptocurrency markets
  • Users managing watchlists and portfolios
  • Users researching potential trading opportunities
  • Users learning technical and market-analysis concepts
  • Risk-conscious traders seeking structured decision support
  • Mobile users requiring personalized market alerts
  • Product operations and compliance teams

Selected product decisions

Position the product as market research and decision support rather than guaranteed financial advice

Keep all final trading decisions under user control

Require explicit confirmation before any consequential action

Use retrieval-augmented generation with current and approved data sources

Display data timestamps and freshness indicators with market-sensitive responses

Separate verified market facts from AI interpretation and scenario analysis

Use watchlist and portfolio context only when authorized and relevant

Explain the evidence and indicators contributing to each insight

Communicate confidence, uncertainty, assumptions, and limitations clearly

Avoid presenting price forecasts as certain outcomes

Support scenario-based analysis instead of one-directional buy or sell commands

Provide educational explanations for technical and market terminology

Allow users to configure alerts without automatically executing trades

Apply stronger safeguards to leverage, derivatives, and other high-risk topics

Maintain audit logs for data retrieval, AI outputs, user confirmations, and system actions

Avoid exposing proprietary prompts, models, APIs, trading logic, or production architecture in portfolio materials

Outcomes

Product Direction

Established a clear product vision and strategy for an AI-powered cryptocurrency market-intelligence and trading-assistance experience.

Strategy

Trader-Centered Experience

Designed user journeys for market analysis, asset research, portfolio review, watchlists, alerts, and conversational assistance.

UX

Grounded Market Intelligence

Defined an AI approach combining current market data, technical indicators, sentiment, research content, and authorized user context.

AI Strategy

Human-Controlled Decisions

Kept consequential trading decisions under user control through review, disclosure, and confirmation requirements.

Responsible AI

Explainable Insights

Designed responses to distinguish market facts, analytical interpretation, uncertainty, assumptions, and supporting evidence.

Trust

Cross-Functional Alignment

Translated product, trading, AI, data, security, compliance, and UX requirements into a unified solution direction.

Leadership

Portfolio Documentation

Reconstructed sanitized product, UX, AI, architecture, evaluation, and roadmap artifacts without exposing proprietary details.

Confidential

Challenges

  • Maintaining accurate insights in rapidly changing cryptocurrency markets
  • Preventing stale market information from being presented as current
  • Separating factual market data from probabilistic AI interpretation
  • Avoiding hallucinated prices, indicators, news, or portfolio information
  • Explaining uncertainty without making the experience confusing
  • Personalizing insights without encouraging excessive risk-taking
  • Combining market, technical, sentiment, news, and portfolio signals consistently
  • Handling conflicting signals across different data sources
  • Preventing users from interpreting AI-generated analysis as guaranteed financial advice
  • Supporting both beginner and experienced traders
  • Designing safeguards for leverage, derivatives, and high-volatility assets
  • Managing latency when retrieving multiple real-time data sources
  • Protecting sensitive portfolio, transaction, and account information
  • Maintaining traceability between generated claims and supporting evidence
  • Evaluating response quality when market outcomes are inherently uncertain
  • Reconstructing the project without exposing confidential implementation details

Lessons learned

  • A trading assistant must help users reason about information rather than replace their judgment
  • Real-time data freshness is a core product requirement rather than only a technical concern
  • Market facts, model interpretation, and possible scenarios must be clearly separated
  • An AI-generated market outlook should communicate uncertainty and alternative outcomes
  • Personalization should improve relevance without increasing pressure to trade
  • Technical indicators are more useful when explained in plain language and combined with context
  • A strong response requires evidence, timestamps, assumptions, and traceable sources
  • High-risk trading topics require stronger disclosures and user-control mechanisms
  • User confirmation should be required before consequential financial actions
  • Product evaluation should measure factual grounding and decision quality, not whether a prediction happened to be correct
  • Portfolio information requires strict authentication, authorization, and data-minimization controls
  • Confidential projects can demonstrate strong product leadership through sanitized and accurately labeled artifacts

Project artifacts

Selected documents and supporting product work.

All

1 artifact
Essential Product Documents Portfolio Edition preview
Public

Document

Essential Product Documents Portfolio Edition

A polished, NDA-safe portfolio document set covering product strategy, user experience, requirements, conversational workflows, market-data integrations, AI architecture, safety controls, evaluation metrics, roadmap, and human-controlled decision support for an AI-powered cryptocurrency trading assistant.

View document