AI Pushes Into Professional Advice

AI FIDUCIARIES

AI to Offer Professional Advice

The weird part of AI is that AI was trained without rules.

It wasn’t told what to do or how to figure things out, it just looked at patterns

—billions, maybe trillions of them. And in certain subject matters, pattern matching is more effective than rigid, rules-based systems.

This has made AI uncomfortable for professionals. We thought some jobs were “safe” from automation, like lawyers or financial advisors. It’s easier to wrap your head around a robot assembling car parts than drafting an estate plan.

This line between what machines used to do and what they will do is blurring.

It’s been less than three years since ChatGPT was released widely, and the technology is used by big banks.

All the major U.S. banks are deploying generative AI.

• JPMorgan is rebuilding workflows

• Citi is revamping wealth management

• Morgan Stanley is modernizing legacy code

• Bank of America is automating client interactions

This rapid adoption is being treated as a potential job killer, like this New York Post headline on Goldman releasing AI firmwide.

This is the unsettling part of AI. We’ve long seen human judgment as beyond the reach of the machines, but it turns out this is (mostly) a bundle of repeatable patterns. (For the record, I think AI changes how work gets done, but that doesn’t mean new kinds of jobs won’t emerge.)

Already, startups like Harvey, backed by OpenAI’s fund, are handling legal tasks for elite law firms. It just raised $300 million at a $5 billion valuation.

AI FIDUCIARY

MIT economist Andrew Lo is going a step further.

He’s developing AI software that could pass the regulatory thresholds required to give financial advice, meeting the fiduciary standard.

Unlike robo-advisors that offer generic allocations, Lo’s AI-powered financial advisor aims to meet the fiduciary duty standard—advice tailored to your best interest, not just your risk profile. It’s a system designed to respond dynamically to individual financial circumstances, not automate boilerplate plans.

Asked recently whether ChatGPT should plan your retirement, Lo said: “Not yet.” The fact that a leading economist even has to qualify his answer suggests we're closer than most people think. (He discusses this at about 5:10 in the interview below.)

Takeaway: AI is moving from automating simple tasks to handling professional judgment in banking and law, and we're closer than expected to AI that could legally manage your money.

SURVEY

Help Benchmark AI Adoption in Finance

It’s hard to get reliable data on how AI is actually being used by financial firms.

Neudata, a data scouting service for the financial services industry, is running a short survey on how the sector works with alternative data, market data and AI.

It takes about 5 minutes. Participants gain access to the results from data buyers as well as vendors, plus a chance to win a $500 gift card. 

Topics include data budgets, in-demand data categories, and the AI models used to process that data. 

Click here for the survey 

ADOPTION

An AI Operating System for Wall Street

The market is full of AI tools—my running list now tracks more than 70 vendors.

While these tools are often pitched for specific tasks, there’s no easy way for them to interact with each other, so you end up doing a lot of copying and pasting between them.

Jason Nabi, founder of WealthAi, is trying to fix that with an AI operating system for finance.

The AI assistant learns how individual users work and becomes “an extension of the user,” adapting to each person's role and preferences, according to Nabi, whose career spans Bloomberg, IBM, and various FinTech startups across London, New York, and Hong Kong.

“We've got this assistant which takes on an identity that is you,” said Nabi from London. “If you're a portfolio manager, relationship manager, analyst, operations manager - it understands you and how you do things, it becomes an extension of you that connects to our platform.”

WealthAi is structured around three main components

1. Provide a dedicated agent front end - Create a personalized AI assistant that understands each user's role and workflow, becoming an extension of them

2. Connect existing legacy tools and systems - Bridge their current vendor packages, data, and internal systems to new AI-driven processes;

3. Create a marketplace of AI apps and service providers - Offer access to both AI applications and traditional service providers through one platform

Since launching from stealth two months ago, they've signed a handful of clients like boutique asset managers, family offices, and private banks. One client decided not to hire additional operations staff and will use the AI assistant instead. Another can now service charities with institutional-level portfolio management at a lower cost. Another is using it for company-wide compliance monitoring.

I think the above examples are indicative of what’s shaking out across the industry—skipping hires, expanding services— it shows that AI is creating and eliminating different tasks. This should accelerate since the cost of using AI has dropped 90% over the last two years and it continues to decline.

"It used to be the case that if you're a family office with less than a couple hundred million [in assets], it's a real struggle," Nabi said. Now smaller operations can access tools that were previously only available to massive firms.

Takeaway:

We're still in the 'dial-up internet' phase of AI in finance. The tools exist, but the infrastructure to make them truly useful for non-technical buyers is just being built now.

PODCAST

Sentiment analysis has been on Wall Street for decades, but I just learned that it generally categorizes news as positive, neutral, or negative, which, when you think about it, is kind of wild. All the headlines, all the news, all the chaos of the world, reduced to just three categories…?

But, like a lot of things we cover in this newsletter, AI is changing that.

In this episode of the Alpha Intelligence Podcast, I talk with my co-host, Francesco Fabozzi—just him, for the first time—about his research into more nuanced sentiment scoring. Instead of boiling everything down to three categories, we explore how AI can assign more granular sentiment values that better reflect market signals.

🎧️ Listen: YouTube, Spotify

NEWS ROUNDUP

Banks, Big Tech Seek AI adoption Guidelines

Citi and Morgan Stanley joined forces with AWS, Microsoft and Google Cloud to set open-source controls for secure AI adoption in the financial industry. FINOS has launched an industry-wide initiative to create vendor-neutral AI adoption standards and real-time compliance tools for financial services, building on its earlier AI risk mitigation work. (CIO Dive)

Quinn raises $11m for AI-powered financial advice

AI-driven financial planning and advice platform Quinn has emerged from stealth with $11 million in seed funding led by Viola Fintech. (Finextra)

Goldman, Citi back AI-Powered Financial Advisor

Conquest Planning, a wealth advisory firm for retail and HNW investors, has raised $80 million in growth equity backed by a host of venture investors from major US banks. (Finextra)

Financial deepfake scams targeted in Senate bill

New legislation seeks the creation of a Treasury-led task force to examine and combat AI-fueled scams that trick Americans out of their money. (Cyberscoop)

QUOTABLE

Billionaire Marks Used AI to Help Write Investing Memo

CALENDAR

Upcoming AI + Finance Conferences

I’ve put together a calendar of upcoming AI and finance conferences. Click on the image below to access links and event descriptions. Let me know if I’ve missed any and I’ll add them. [email protected] Thanks!

RESOURCES

If you're getting up to speed on AI, here are a couple of resources that are worthwhile:

One is from the professor Ethan Mollick, who I think has pretty measured takes on what's capable and not capable with AI, and tends to focus more on the practical side of things.

Anthropic also has a free course on understanding AI:

WHAT ELSE I’M READING

• Advances in AI will boost productivity, living standards over time (Dallas Fed)

• AI Hype Is Proving to Be a Solow's Paradox (Bloomberg Opinion)

• Monitoring Portfolio News Using AI (QuantStreet)

• Four months after a $3B valuation, Harvey AI grows to $5B (TechCrunch)

• AI Agents Make Humans More Productive, MIT Researchers Say (Bloomberg Law)

• A teenage podcaster shares 4 life lessons from Steve Cohen, Howard Marks, and other Wall Street giants (Business Insider)

• Citadel lures another Man Group ArcticDB engineer (HedgeWeek)

• How BlackRock Built AI Agents for Asset Management with LangGraph