How AI revolutionized investment decisions at a major asset manager

I recently spoke with a client at a major asset manager about their internal AI initiative and how it has reshaped its business. Since the discussions were confidential, I am not going to disclose the name of the firm but what he told me was fascinating.

In 2023, the firm launched its “Intelligent Investing Initiative,” embedding AI across its asset management operations. Just a year later, it has completely transformed the business. Here’s how it unfolded:

The company’s ‘MarketMind’ AI, trained on decades of data and real-time satellite imagery, predicted the global cocoa price surge several months early by tracking cocoa planting and weather in Cote d’Ivoire. The firm’s tech fund avoided $4bn in losses by shorting overvalued AI startups before they announced profit warnings. However, not all forecasts were perfect: A false positive about a cyberattack on JP Morgan briefly crashed the AI’s credibility, until human analysts overruled it.

Professional fund managers now can mimic other star managers’ portfolios in real-time. One fund manager uses MarketMind to replicate Bridgewater’s “All Weather” strategy.

The tool also curbed emotional decisions: When some portfolio holdings started to sell off, the AI deployed calming memes and historical rebound data on social media platforms. This created additional inflows from buy-the-dip retail investors and reduced the drawdown by an estimated 45%.

The company is increasingly using client data to improve the performance of its investments. By 2024, the company’s AI platform analysed data from 50+ sources of clients – bank accounts, wearables, and even LinkedIn career trajectories – to adjust the holdings of fund managers.

Compliance is also increasingly outsourced to the algorithm. The company’s AI-powered compliance system slashed regulatory fines by 90% by flagging insider trading patterns in Bloomberg IB chats. However, an internal scandal erupted when the AI greenlit investments in a cobalt mine accused of child labour, having misclassified “artisanal mining” as ESG-friendly.

Indeed, fund managers are increasingly being made redundant. Gone are the days of human fund managers squinting at earnings reports or debating macroeconomic trends. Instead, the firm now employs ‘AI Whisperers’ (ex-analysts who translate model outputs) and ‘Quantum Ethicists’.

This has led to the predicted backlash, which, unfortunately, doesn’t seem to be welcome by management. During a town hall, when a VP asked why the AI kept investing in companies with negative EBITDA, the CIO snapped, “EBITDA is a social construct.” Employees now receive mandatory training titled ‘Embracing the Black Box: How to Trust Machines’.

So far, the company’s AI-managed funds outperformed human-led peers by 3.2% annually, helping to attract new assets.

Yet debates persist: Can AI navigate moral trade-offs (e.g., maximizing returns vs. avoiding authoritarian regimes)? Does mimicking star analyst portfolios increase market risk? And why does the AI still invest in Tesla every time Elon Musk tweets a rocket emoji?