How AI Is Changing Investment: BlackRock's David Hefter on Quant Models for Qualitative Data

BlackRock AI investment lead David Hefter in conversation with Mo Salah shared critical insights around AI modernizing investments, ripple effects for broader markets, and actionable advice for founders.

Many many gems, I am selecting this one by David: "AI allows quantitative models to be applied to qualitative data"

This resonates: I built an AI stock-prediction product 20 years ago, early expertise that has only compounded* since! :-) (compounded = iykyk)

What is the world’s largest asset manager seeing and what it means for you!

BlackRock's scale

BlackRock now manages a record $14.04 Trillion in assets. (150X bigger than the biggest VC funds)

BlackRock’s Internal AI Strategy

The core of BlackRock's strategy isn't just "using AI", it is about becoming smarter investors by applying quantitative rigor to qualitative data.

In David Hefter's words "AI allows quantitative models to be applied to qualitative data"

The Shift: Traditionally, investment decision-making was document-heavy; investors read endlessly.

The New Flow: AI allows BlackRock to "systematize textual data." By using AI as a deep research tool and augmenting it with proprietary internal data, they can now apply quantitative models to qualitative sources (news, filings, reports).

Data is Key: Hefter noted that model capabilities are becoming commoditized ("not that different from one model to another").

The true alpha lies in the data, not the model.

The Future: Smaller, Reasoning Models

Echoing the theory endorsed by Prof. Jason Corso, the future of AI in finance is not necessarily "bigger is better."

Reasoning > Knowledge: Models do not need to "bake in" all the world's knowledge, they can be smaller, efficient that reason on cue, retrieving the specific knowledge they need when they need it.

The "Supply Chain" of Information

A classic investment example is analyzing how a supply chain is affected by disparate factors.

Integration: Critical data is often trapped in disconnected silos: a podcast, a public filing, a news article.

The Unlock: AI enables the integration of trapped data points, systematizing them into a single, coherent view that human decision makers might miss if they relied on manual reading alone.

Advice for Founders

David Hefter offered two distinct pieces of advice for founders.

Be Differentiated: You/Your company must stand out. (Note: Me being an AI pioneer who built products 2 decades ago and being a Polymath and our team :-).

Disrupt or Facilitate: The market is full of massive legacy entities. Make a clear strategic choice: are you building to disrupt them, or are you building to facilitate them?

I will be sharing more insights in separate articles (LinkedIn word limit) from Chris Yeh (Blitzscaling Ventures, bookmarks for authors :-), Sydney Sykes (NVIDIA), Andy Tsao (SVB hosted), Amy Benziger (The Aspen Institute), Lynda Chin, Zarko Ackovik (DWTC), Kevin Jiang (Mangusta Capital), Mariane Bekker, and I missed many!