The wealth management industry has spent two years captivated by generative AI — chatbots, summaries, draft emails. Then on February 10th, Altruist launched AI-powered tax planning inside Hazel, its $60-per-month platform, and Schwab, LPL Financial, and Raymond James each dropped between 7 and 11 percent in the sessions that followed.
The industry noticed. The question is whether it understood what it was seeing.
The distinction between generative AI and agentic AI is not academic. It is the difference between a system that answers questions and one that takes action — and for advisory firms, it is the dividing line between AI as a productivity tool and AI as an operating system.
What Generative AI Gave Advisor
Speed at things they already did. Drafting emails. Summarizing notes. Useful, certainly. But Kitces Research found advisors spend just 20 percent of their time actually meeting with clients. Generative AI made the other 80 percent faster. It did not eliminate it.
What Agentic AI Changes
Agentic AI does not wait for a prompt. It perceives conditions, reasons through multi-step workflows, and executes — monitoring tax exposures, updating client profiles, flagging compliance gaps. The advisor shifts from operator to decision-maker.
This matters because McKinsey projects a shortfall of 100,000 advisors by 2034, with 110,000 expected to retire this decade and just 8,000 net new recruited in the last ten years. That gap will not be closed by hiring. It will only be closed by infrastructure.
The Infrastructure Question
Firms that adopted generative AI first will not necessarily lead in the agentic era. Generating a summary requires language models. Executing workflows across custodians, compliance engines, and client records requires orchestration, memory, and continuous intelligence.
This is the distinction WealthDirector was designed around — not AI that responds to advisors, but an intelligence layer that continuously monitors client-portfolio alignment and acts across the full wealth lifecycle.
The question is no longer whether to use AI. It is whether you have the infrastructure for AI that acts.