Real-World Case Studies: AI in Investment Management
From the course AI for Investment Analysts
Major asset managers and hedge funds have publicly disclosed AI deployments that have changed how research and portfolio management work.
Case Study 1: BlackRock Aladdin — AI-Powered Risk at Scale
BlackRock's Aladdin platform manages risk analytics for over $20 trillion in assets under management across BlackRock and external clients.
AI Role: Aladdin ingests real-time pricing, news, economic data, and alternative data to run scenario analysis and stress tests across millions of positions simultaneously. Machine learning models flag correlation breakdowns and tail-risk events faster than any human team could monitor.
Outcomes:
- Real-time portfolio risk monitoring across asset classes and geographies
- Faster identification of liquidity mismatches during market stress
- Standardised risk language used by hundreds of institutional investors
Key Lesson: AI-driven risk infrastructure creates a durable competitive moat. The barrier is not the algorithm — it is the proprietary data and the decades of parameter tuning Aladdin has accumulated.
Case Study 2: Two Sigma — Data-Driven Systematic Investing
Two Sigma manages over $60 billion using quantitative strategies built on machine learning and alternative data.
AI Role: Two Sigma ingests satellite data, shipping records, credit-card transaction datasets, and social media sentiment to generate trading signals that human analysts cannot see or act on at the required speed.
Outcomes:
- Consistently strong risk-adjusted returns across market regimes
- Ability to trade strategies across thousands of instruments simultaneously
- Faster signal decay detection and model refresh cycles
Key Lesson: At the frontier, the alpha is in the data sourcing and signal hygiene, not the model architecture. Two Sigma's edge comes from proprietary data cleaning pipelines as much as from AI models themselves.
Case Study 3: Goldman Sachs GS AI Platform — Analyst Productivity
Goldman Sachs deployed internal AI tools to its research and investment banking divisions, publicly stating that AI can perform tasks that previously required first-year analyst cohorts.
AI Role: Internal LLMs draft first-pass equity research summaries, generate financial model assumptions from earnings transcripts, and produce due diligence document reviews for M&A transactions.
Outcomes:
- Reported reduction in time for initial research drafts
- First-year analysts redeployed toward higher-judgment tasks
- Standardised formatting and compliance language across research output
Key Lesson: AI does not eliminate analyst headcount at Goldman — it raises the quality floor of what junior analysts produce on day one, making the entire team more productive.
Lessons for Independent Analysts
- You do not need institutional infrastructure to use AI; ChatGPT Code Interpreter replicates many Aladdin-style scenario functions for individual portfolios
- Alternative data that was once available only to hedge funds (satellite imagery, credit card data) is now available via API from vendors at accessible price points
- The competitive gap is narrowing: AI democratises quantitative analysis that once required a team of PhDs
Key Takeaways
- •BlackRock Aladdin shows AI's value is in data accumulation and parameter tuning, not just algorithms
- •Two Sigma's edge comes from proprietary data cleaning pipelines as much as from ML model quality
- •Goldman Sachs redeploys analysts toward higher-judgment work rather than eliminating them
- •Alternative data that once required institutional resources is now available to independent analysts via API
Progress is kept in this browser only.