Anthropic launches Claude for financial advisors and takes on OpenAI in the race for finance's paperwork

By PYMNTS · September 14, 2026.

Anthropic announced on Monday, in a post on its corporate blog, the launch of "Claude for Financial Advisors," a package of connectors and "skills" (specialized work modules) designed to automate meeting preparation and paperwork for financial advisors. According to the company, the tool connects Claude to custodians (the entities that safeguard clients' assets), portfolio management platforms, CRM (Customer Relationship Management, the software where the relationship with each client is recorded) systems and the planning tools advisory firms already use.

With those systems connected, the new skills allow Claude to prepare meetings, analyze portfolios and carry out regulatory compliance reviews. Anthropic stresses that the advisor keeps control: critical tasks require their approval, and a specific compliance skill reviews client-facing language to spot problems, as well as helping firms document their reviews and their use of AI in line with the rules of the SEC (Securities and Exchange Commission, the U.S. stock market regulator).

"When Claude handles administrative tasks such as meeting preparation, follow-up drafts and prospect pipeline review, advisors can spend more time serving clients," Anthropic said on its blog. The company argues that freeing up that time would allow firms to expand their client base and offer more tax, estate, insurance and retirement planning services; that is the company's own projection, not an already measured result.

This is not Anthropic's first foray into the sector. In May it had already unveiled ten AI agents for financial services aimed at tasks such as producing pitchbooks (the presentation dossiers for a deal) and KYC (Know Your Customer, the process of verifying a client's identity and background) checks. Claude for Financial Advisors extends that bet to the more retail segment of advisory work.

The move comes just four days after OpenAI announced, on September 10, its own tailored offering for the sector: ChatGPT for Financial Services, focused initially on investment banking and equity research, designed to help generate reports, financial models and client materials. And on September 11, PYMNTS reported that the startups Rogo and Hebbia —which turn written instructions into financial models, memos and presentations by connecting to the data each bank has under contract and to their own files— run, precisely, on models from Anthropic, OpenAI and Google.

That last detail is the one that says the most about where this fight is heading. Anthropic sells the same model that powers direct competitors of its own vertical product, and at the same time competes with part of its own API customer base. It is not a contradiction: it is the standard strategy of an infrastructure provider that also wants to own the application layer. The real dispute is no longer so much about which model reasons better —that debate remains open— as about who controls the integration with the custodian, the CRM and each sector's daily workflow: the war is being fought in the plumbing, not on a benchmark scoreboard.

For the ordinary financial advisor, the message fits what we had been observing in banking: administrative work gets automated first —preparing meetings, drafting follow-ups, reviewing the prospect pipeline— while judgment and the client relationship hold up and, if Anthropic's design works as promised, are even reinforced. The employment risk falls not so much on the advisor with their own book of clients as on support staff —junior analysts, paraplanners, administrative workers— whose role consisted above all of producing that paperwork.

The most significant thing in the medium term may be the design of the compliance piece, not the productivity one. Building into the product a skill that documents the use of AI in line with SEC rules acknowledges something that other regulated sectors —healthcare, law— will have to accept sooner or later: there, AI is not adopted unless it comes with an auditable trail and a human to sign off. And if it really frees up hours so that firms can serve more clients with tax, wealth or retirement planning, it is a small but concrete example of our underlying thesis: AI does not replace expert judgment, it makes the time needed to exercise it cheaper, and that —well managed— is a first step toward quality financial advice within reach of more people than just the high-net-worth clientele who can afford it today.