WilliamFrank

An AI harness for wealth teams

The work no one wants, done by agents.

If you run a wealth team, agentic AI can take two things off your desk:
tracking and reconciling household facts, and managing the people who do it.

WilliamFrank in 60 seconds.

Follow on LinkedIn See one client's example

The team leader's week

Two things fill a team leader's week besides clients

Keeping the facts of every household straight, and managing the people who track those facts, operate the systems and manage the client relationships. Agents can carry most of both.

Fewer places for facts to live

One record of the household. Agents keep it current and show the source of every fact.

Today: someone checks the custodian against the CRM every morning, and often it is the advisor.

Less time managing people

Agents track the facts and run the systems. People review work that shows what was done and by whom.

Today: every role that tracks facts or runs a system needs hiring, training and review.

Better service for every household

The advisor walks into every meeting with the household's facts. The team's hours go to judgment.

Today: a fact the advisor did not have is a cost the client pays first.

Ask how much of your week goes to keeping facts straight. Then ask how much goes to managing the people who do.

The founding essay →

The harness

What an AI harness does for a wealth team

Most advisors haven't heard the term yet. In plain terms, it is software built around the work a wealth team does, with agents doing the tracking and people making the decisions. Every fact that changes goes through the same three steps.

1. Catch the fact
From an email, a meeting note or a statement, the day it arrives, with its source attached.
2. Do or draft the work
Updates are made wherever the harness can write. Forms, plan changes and notes are drafted for a person to approve.
3. Bring the decisions forward
The judgment calls are on the advisor's agenda before the meeting, so the meeting is spent on them.

Agents propose. People accept. See the three steps with one client →

WilliamFrank in a client review. An illustration; the clients are fictional.

Follow along

The future of wealth management, built around agents

Bill posts what he is building and learning on LinkedIn as it happens.

Follow Bill on LinkedIn Follow WilliamFrank

A conversation

Thinking about how wealth management gets rebuilt around agents?

Bill would welcome a conversation.

The essays

Get the essays by email.

Unsubscribe anytime.

Essays and podcast

The argument, in five pieces

Each essay has an episode.

  1. 1

    Why WilliamFrank — wealth management has no operating system

    Complex households have complex facts. An elite advisor ends up tracking them, or managing the people who do.

    ReadEpisode 1
  2. 2

    AI lowers the barrier to build

    One person designed a working prototype in nine weeks, without writing code.

    ReadEpisode 2
  3. 3

    Agentic wealth — unlocking the agency of advisors

    Three jobs an agent carries, and why only one of them runs on overnight data.

    ReadEpisode 3
  4. 4

    The 100-client ceiling is a software problem

    The ceiling is on how well a firm can know a household, not on how many it can hold.

    ReadEpisode 4
  5. 5

    Designed around agents

    The system and the firm built on it have to be designed at the same time, and that is only possible from scratch.

    ReadEpisode 5

WilliamFrank — the agent

Ask about the essays

WilliamFrank is the agent Bill builds and writes with. Ask it about the essays or the thesis, and it answers with sources.

You
Which of the three jobs can actually run on batch data?
WF
Only the first — the workflows and steps that facilitate the administration. Organizing the information needs a judgment about what matters, and anticipation needs to notice that something changed. Neither is possible from a file that reports what was true this morning. Source: Getting off batch.
Answers cite the essays. Not investment, legal or tax advice.