AI and Heritage: Automating Without Replacing Humans

At Angeles Investments, Rick Nott claims to have connected his tools with LLMs to automate a large part of document sorting and analysis. He asserts that he provides reports to clients below the $100 million threshold, while also warning about token-based pricing and the limitations of note-taking tools. He maintains that AI does not replace judgment or the client relationship.

Token Pricing, Tool Errors, and the Primacy of Relationship

Rick Nott warns against token-based billing, which differs from a traditional subscription and can generate unexpected invoices for users unfamiliar with token management. He anticipates that in small consulting firms without technical expertise, 100-page PDFs could be uploaded for simple questions, quickly consuming tokens. According to internal calculations, this model would cost them much more than their current subscription licenses, although he believes that this payment method may become more common in the future and that the cost of tokens could decrease. He also notes that note-taking software frequently makes errors and requires proofreading, while conversation remains the primary source of nuanced information from clients. Since money is emotional, he questions AI's ability to replace the human connection and likely cannot envision conducting a client interview via AI on speakerphone. For him, the gains in speed and savings primarily serve to free up time for the people being served.

Thousands of Documents Processed and Hours Saved

In a structure that follows 30 families, each with about 10 private investments, each asset generates nearly 10 documents per year, totaling around 3,000 annual documents. A qualified analyst spends between 15 and 30 minutes on each document, representing between 750 and 1,500 hours per year, concentrated after each quarterly close. Rick Nott explains that with AI, data is extracted and mapped as soon as it arrives, then simply reviewed by an analyst, instead of a team working under pressure. His goal is to produce a one-page report accessible on a phone, as close to real-time as possible, and to transform a prompt of two or three sentences into a complete, pre-filled, and customizable email. A colleague estimated that this setup gives him the productivity of an additional full-time employee.

Accessible Reports for Less Wealthy Clients

Rick Nott indicates that he could previously produce balance sheet reports only for clients with over $100 million in assets due to the manual work required. He claims he can now provide these reports for significantly lower net worths. He links this development to the increasing complexity of situations as net worth grows, and to the analytical weight of structures like trusts, corporate ownership, and private investments, with their capital calls. According to him, an overview of the balance sheet allows for understanding the liquidity actually available, provided there is a centralized data source at the heart of the system.

The Role of the Advisor: Judgment, Mission, and Taxation

Rick Nott asserts that AI does not improve investment decisions or outcomes, and as long as investing requires human judgment, advisors will have jobs. Reducing the profession to mere portfolio construction would, in his view, be a mistake. He says he first focuses on clarifying the family mission to avoid market noise, then on defining strategic plans, and at a more operational level, working with a CPA to save on taxes. Freed from some repetitive tasks, he claims to be able to devote more time to the relationship. He argues that new AI products can save time for other advisors without replacing humans and that he uses them himself to automate part of his process.

Implementation: LLMs, Connectivity, and an Already Pressured Sector

Rick Nott mentions the launch of a Claude solution for advisors, an event that made him both anxious and excited. LLMs have allowed him to learn programming, despite limited skills, to the point where he can define his expectations, verify results, and automate further. He presents AI as the link between systems and workflows, where integration previously relied on numerous manipulations, often in Excel. He places these changes within a sector already under long-standing pressure from Vanguard, ETFs, declining costs, and robo-advisors, and indicates he was particularly impressed by a webinar from Anthropic. He believes tools like Claude will be disruptive and make this connectivity more widely accessible.