AI for Asset Management: Claude Tools for Portfolio and Fund Management
How institutional asset managers use Claude for portfolio construction, factor analysis, performance attribution (Brinson-Hood-Beebower), manager selection, and regulatory reporting.
Asset Management and AI
Asset management combines rigorous quantitative analysis with investment judgment. Claude with ClaudeFinLab accelerates the quantitative layers — portfolio construction, factor analysis, performance attribution, and client reporting — so portfolio managers can focus on alpha generation and client relationships.
Portfolio Construction and Optimization
- "Build a mean-variance efficient frontier for this 12-asset portfolio. Input: expected returns, covariance matrix (attached). Find the minimum variance portfolio, maximum Sharpe ratio portfolio, and the portfolio with 10% target volatility. Report weights, expected return, volatility, and Sharpe for each."
- "Apply Black-Litterman optimization. Market cap weights (attached). My views: (1) US large-cap growth outperforms blend by +2%/yr with 60% confidence; (2) Emerging markets outperform developed ex-US by +3%/yr with 50% confidence. Compute BL posterior returns and optimal BL weights."
- "Analyze this portfolio's factor exposures using the Fama-French 5-factor model: market beta, SMB (size), HML (value), RMW (profitability), CMA (investment). Show trailing 3-year factor loadings and t-statistics. Which factors explain the most return variance?"
Performance Attribution
Understanding where returns came from — allocation, selection, or interaction — is core to investment management:
- "Perform Brinson-Hood-Beebower attribution for Q2. Portfolio return +4.8%, benchmark +3.9%. Sector weights and returns (attached for portfolio and benchmark). Decompose: allocation effect (over/underweight sectors), selection effect (stock selection within sector), interaction effect."
- "Compute risk-adjusted performance metrics for the past 3 years: Sharpe ratio, Sortino ratio (downside deviation), Calmar ratio (max drawdown), Information ratio vs benchmark, Treynor ratio. Annualize all metrics."
- "Analyze this fund's rolling 12-month alpha vs the S&P 500. Has alpha been persistent or episodic? What market regimes (rising rates, falling rates, high vol, low vol) coincide with positive and negative alpha periods?"
Factor and Style Analysis
- "Run a returns-based style analysis (RBSA) for this fund using 8 Sharpe style boxes (large/mid/small × value/blend/growth, plus bonds). Use 36-month rolling regression. What is the style drift over time?"
- "Compute the portfolio's current factor tilts vs the Russell 1000: P/E ratio, P/B ratio, dividend yield, earnings growth (3yr CAGR), revenue growth, ROE, beta. What style risk does the current portfolio carry?"
- "Screen the S&P 500 for quality factor: define quality as ROE > 15%, debt/equity < 0.5x, earnings quality (accruals ratio < 0.05), and EPS stability (coefficient of variation < 25% over 5 years). How many names pass all screens?"
Manager Selection and Due Diligence
- "Evaluate this long-only equity manager: 5-year track record vs Russell 1000, 85% correlation to benchmark, active share 62%, average holding 47 names, turnover 42% annually. Compute information ratio, tracking error, active risk. Is this a high-conviction active manager or a closet indexer?"
- "Compare these two fixed income managers: Manager A — 7yr duration, 4.2% yield, IG only; Manager B — 5yr duration, 5.8% yield, IG/HY blend. Compute expected return per unit of duration risk and per unit of credit spread duration."
Client Reporting and Suitability
- "Draft a quarterly client commentary for a balanced portfolio (60/40). Portfolio returned +5.2% vs 60/40 benchmark +4.7%. Key drivers: overweight to US large-cap growth (+1.1% contribution), underweight to international (+0.3% detractor), duration shorter than benchmark (+0.4% contribution). Tone: informative, confident, 300 words."
- "Assess suitability for a 58-year-old client: $2.4M portfolio, 80% equity/20% fixed, moderate risk tolerance, 8-year investment horizon, retirement in 7 years, pension income $36K/year. Does the current allocation match the risk profile? What glide path would you recommend?"
Regulatory Compliance
- "Review this investment mandate for MiFID II best execution compliance: the fund's order routing policy. Does it address: price, costs, speed, likelihood of execution, likelihood of settlement, size, nature, and other considerations? Flag any gaps."
- "Generate the required MiFID II cost and charges disclosure for this fund: ongoing charges 0.75% TER, entry charge 0%, exit charge 0.5% (first 12 months), performance fee 20% above high water mark. Format as a standardized PRIIP KID cost table."
Fiduciary note: Asset managers have fiduciary duties to clients under the Investment Advisers Act (US) and MiFID II (EU). AI tools produce analysis — final investment decisions remain the responsibility of the portfolio manager. All AI-generated output must be reviewed before client delivery.
Recommended MCP Setup
{
"mcpServers": {
"claudefinlab-portfolio": {
"url": "https://claudefinancelab.com/portfolio/sse",
"headers": { "Authorization": "Bearer YOUR_API_KEY" }
},
"claudefinlab-market": {
"url": "https://claudefinancelab.com/market/sse",
"headers": { "Authorization": "Bearer YOUR_API_KEY" }
}
}
}