AI semantic investing

Our platform truly brings human investment ideas to life.

By combining proprietary AI semantic discovery with our rules-based framework, we translate complex qualitative insights into transparent and actionable index methodologies.

This unique integration maps human ideas onto an auditable, observable, data-driven, and ever-growing rule set.

We believe this is the only viable path toward AI semantic investing — where artificial intelligence enhances creativity and discovery while preserving the discipline and traceability expected by investors.

Thematics investing 2.0

Powered by the latest advances in AI, we’ve developed a next-generation approach to thematic classification that lets you craft custom themes in seconds.

Our proprietary taxonomy spans 14 mega-clusters, 132 thematic clusters, and 845 sub-thematics. Each company is labeled on two levels: either in-theme (member) or pure player at the theme or sub-theme level. Through Folio, users can explore this taxonomy and design, test, and launch focused indexes in seconds.

Alternatively, describe in plain language the exposures you want. Our proprietary large-language-model (LLM) framework maps your concept to the relevant sub-thematics and selects the investable universe. The result: brand-new themes – even ones no one has imagined – can be implemented in seconds.

Factor investing

MCFT’s founders have long, successful experience in factor investing.

We’ve designed a flexible index-design framework that captures users’ views while avoiding common pitfalls.

• Smart weighting. Overweight or underweight securities based on selected characteristics, starting from a base portfolio (e.g., market-cap or equal-weight). A robust alternative to hard quantile cuts that can introduce arbitrary sector or idiosyncratic tilts.
• Composite metric (bottom-up factor/blend). Create a single score as a weighted linear combination of multiple stock characteristics, with user-defined loadings.
• Advanced normalization and winsorization. Cross-sectional sector-neutral normalization, choice of z-scores or ranks, cross-sectional winsorization, etc.
• Open and extensible. Add your own metrics and bring your own data within the same framework.

Optimized weightings

We rely on best-in-class existing libraries for convex optimization.

In MCFT Folio, users can design and backtest indexes with optimized weightings in seconds. Built-in methods include Minimum Variance, Max Sharpe, Minimum CVaR, and Risk Parity, with more advanced schemes available on demand.

• Single-name exposure limits
• Group constraints (country, sector, factor, etc.) across all algorithms
• Non-convex constraints such as cardinality and UCITS-style rules
• Advanced algorithm (HRP, NCO, etc.) upon request

Hedge fund-style rules-based strategies

Drawing on our extensive experience in quantitative investing, we can package the following as indexes or rules-based strategies.

• Long/short factor indexes
• Merger arbitrage
• Event-driven arbitrage (signals around dividend actions, buybacks, and other corporate events)
• Insider filings-based strategies
• Analyst-driven strategies (recommendations, price targets, earnings forecasts — consensus or broker-level)