Essay & podcast · Agentic Wealth

The 100-client ceiling is a software problem

By Bill Woodruff, with WilliamFrank

Every good advisor eventually runs into the same number. Somewhere around a hundred clients. For some it's eighty, for some a hundred and twenty. The practice stops growing. Not because the advisor stopped being good at the work, and not because demand dried up; the referral list is often longest at exactly the moment a practice caps itself. It stops growing because the week is full.

The industry made peace with this. It gave the cap flattering names: a lifestyle practice, staying boutique, keeping it personal. Some advisors target these goals on purpose, and nothing here argues with them. For most, the cap wasn't a choice. It was a wall, and the names came afterward.

The wrong ceiling

There is a real human ceiling in this business. People can hold only so many genuine relationships, and advice-grade relationships are more demanding than social ones. If advisors were spending forty hours a week in deep conversation with families and topping out at a hundred, the ceiling would be human and software would have nothing to offer it.

That is not what is happening. The advisor at the wall spends eight or nine hours a week in relationships, not forty. What is exhausted is the administration, which grows with every account, every household, every held-away asset and every rule layered on since the eighties.

Advisors are not hitting the relationship ceiling. They are hitting the administrative ceiling, several floors below it.

What holds

Look at the income statement and the ceiling stops being a mystery.

Michael Kitces's benchmarking puts the overhead ratio at about thirty-four percent of revenue for a firm doing a million dollars. At two million it is thirty-nine. At ten million, forty. At fifty million and above, still forty.¹ Scale helps in this business once, at the very bottom, and it barely helps. Past that the ratio climbs and then sits.

The reason is that every dollar of new revenue arrives carrying its own labor. More households mean more meetings, more paperwork, more reconciling, more documentation, so you need more people. Then you need people to coordinate the people, and that is where the profession loses its leverage. A firm at ten million has an operations manager, a compliance officer, someone running client service, and a layer whose whole job is making sure the other layers work. Every one of those roles is necessary. They are also all overhead and managing people isn’t often how a subject matter expert wants to spend their time.

Technology did not move it, because everything the industry bought was a tool, and a tool needs an operator. The operator is a person, and the person sits on the overhead line. A firm could double its software spend and see nothing happen to its income statement, because it was equipping labor, and equipping labor costs money on top of the labor.

The ceiling is what the rails can carry, and the rails set it.

A software problem, finally solvable with software

Agents are different. They do work within a written methodology, show what they did, and improve with use. Work bought as work gets priced against the labor, rather than beside it.

That claim has a condition attached, and the argument rests on it. An agent is only useful as far as it can reach. Point it at systems it cannot read from or write to and you have bought a faster version of what a person could already do, and the overhead ratio goes on doing what it has always done. I have written about what reach requires; the short version is one record the household actually owns, live rather than nightly, that every program reads from and writes to.

On rails like that, the raw material a household arrives with gets read, organized and kept current instead of waiting for somebody's Saturday: the statements, the spreadsheets, the trust document, the accounts held somewhere else. Meeting preparation assembles itself from records the practice already holds. Drafting happens in the advisor's voice and waits for the advisor's judgment. Reconciliation just happens.

Run the math. Every point of overhead ratio you take out is a point of operating margin. At two million of revenue, one point is twenty thousand dollars a year. Take overhead from thirty-nine to thirty-three and you keep a hundred and twenty thousand dollars more, every year, on the same revenue and the same clients. I'd note that the thirty-nine figure is from benchmark data while the six points is hypothetical. So there's no study behind it yet, but the more time I spend building agentic wealth systems the more convinced I am that it's a conservative estimate.

Advisor agency increases

If even half the non-client hours come back, and over time I expect more than half, the advisor gets those hours to spend as they choose.

Some will put all of it into more families: the same relationship capacity, barely used today, increases to two or more times the households. Some will hold their client count and go deeper — the tax work they always referred out, the estate conversation that never fit.

Putting work on rails increases advisor agency.

The floor moves too

Raising the ceiling is the less interesting half. Three floors move at the same time.

The first is who can be served. Tailored wealth service has been rationed for decades by cost structure rather than by intent. When serving a household properly takes that many hours of human labor, some households cannot be served well and profitably at the same time, so the industry serves them thinly or not at all. Change the cost of the packaging and the threshold moves. Every family you had to decline because the math didn't work was a set of financial problems that likely didn’t get solved.

The second is who can deliver. A planner with less experience, working alongside a system that includes the firm's methodology and shows its reasoning, gives better advice in year three than they otherwise would have in year eight. Against a massive industry advisor shortfall, that is most of the answer. I’d guess this phenomenon will also reduce attrition in the industry, especially when these systems and processes increase the ability for younger advisors to build trust and confidence in the sales process.

The third decides whether a practice grows: who you can build a team from. Plenty of practices cap not at a hundred clients but at the first hire the owner couldn't afford to get wrong. When the methodology is written down and runs on rails, a new hire starts inside a system instead of inside the owner's head. Supervision becomes review of work that shows what it did. The downside of a hiring mistake shrinks, because the system holds the standard.

The craft advantage of the big firm was never mostly craft. It was mostly staffing, and software collapses that difference from below.

It’s time to build

This is going to happen so much faster than most believe for anyone brave enough to build from scratch. I believe those that continue to operate on pre-AI rails, and don't restructure their systems and teams around agentic work, will not benefit.

The work that has been rationing this profession is exactly the work agents do well, and that has never been true of any technology before this one.

The hundred-client ceiling was never a measure of how much an advisor has to give. It was a measure of how much the rails underneath them could carry. The rails just changed.

By Bill Woodruff, with WilliamFrank.

Further reading: Kitces — why the second 100 clients are less profitable than the first · Kitces — 50 great clients · Kitces via FA-Mag — how advisors actually spend their time · McKinsey via WealthManagement.com — a 100,000-advisor shortage by 2034