Skip to main content
The Financial Times recently asked whether AI could help identify skill in fund managers. The deeper opportunity, however, is not just identifying skill but embedding intelligence across the entire multi-manager lifecycle — from research to pricing, from portfolio construction to performance attribution.

Key Points

1. From Research to Evidence-Based Pricing

In a multi-manager environment, research teams drown in unstructured data: manager notes, due diligence reports, market context, and historical performance metrics.

AI can standardise this toolchain — extracting key factors from past reports, summarising narratives, and running elasticity simulations to test fee sensitivity.

→ Outcome: faster, more consistent research output, and pricing anchored in data rather than negotiation intuition.

2. AI as a Meta-Analyst

Beyond automating notes, AI can learn what signals historically preceded outperformance — qualitative (team stability, philosophy coherence) and quantitative (alpha persistence, drawdown patterns).

This turns the research platform into a meta-analyst that detects emerging skill or risk decay earlier than periodic reviews.

3. Feedback Loops Across the Lifecycle

Each phase — research, selection, monitoring, and pricing — generates intelligence that improves the next.

By closing the loop, AI shifts due diligence from a static report into a living system of skill inference.

→ Example: a manager flagged for inconsistent factor exposures automatically triggers a note update and repricing scenario.

4. Amplifying Human Judgment

AI doesn’t replace manager selection. It gives human committees augmented memory, speed, and analytical reach — ensuring that intuition operates on complete, standardised evidence.

5. Tangible Metrics of Success

In the Gemmo model, success looks like:

  • 2× faster note drafting
  • A single source of truth for the manager pipeline
  • Pricing decisions supported by elasticity simulations and benchmark data
  • Continuous AI-driven skill scoring from live performance feeds

Closing Thought

Where the FT article sees AI as a diagnostic of fund-manager skill, the multi-manager world can go further: AI as an orchestration layer that connects research, selection, and pricing into one adaptive system.
That’s not about predicting bubbles — it’s about building resilience and evidence into how asset owners allocate capital in the first place.