They're Selling You the Wiring and Calling It Electricity

Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.

The vendor’s deck is usually very good. Somewhere around the sixth slide there is a picture of your firm as it exists today, which is unflattering but largely accurate. You have systems that do not speak to each other, client data living in seven or eight places, and more than a dozen institutionalized workarounds that some capable person invented years ago, but nobody is sure anyone ever wrote them down.

That slide describes your firm, and the following slide arrives carrying the remedy. Unify the data, buy the platform, hire the agents, and you are AI-ready.

The vendors are right about the problem, and if you are a regular reader of my work, you know their problem statement is also my own. Most companies in this industry still run on systems that do not talk to each other, on institutional knowledge that lives in three or four people's heads, and on processes that vary by practice because nobody ever decided which version was correct.

Artificial intelligence laid over fragmented data produces noise. I argued in “The AI Fortress” that intelligence can counter the drag, but only on a clean operating model, never on one you keep trying to fix by adding to it. What follows takes that argument a step further, for the owner who has to live inside the choice.

The trap on the seventh slide springs easily, and the pivot from the sixth slide to the seventh deserves a harder look than most leaders give it. Notice that nothing in the diagnosis (nothing is standardized and machines need standardization) actually implies the conclusion (buy our software).

The deck has established that your operating model is illegible and then proposed a remedy that does not make it legible. A platform selling into an illegible operating model has no commercial reason to make the model legible, since the illegibility is the market, and a market that pays better to work around friction than to remove it will reliably produce the products we are all being sold.

Where the Sleight of Hand Lives

The easy version of this conversation points at the appliance, the notetaker, the chatbot, the meeting capture, or the thing on the screen that impressed somebody on your leadership team. The vendors know better than that now, which is why their decks improved this year, and they will tell you the appliance is useless without something underneath it to carry the power. They are right about that, too.

Many have stopped selling the appliances and are now selling the wiring (a unified data layer, conduit, panels, and a path for the current to travel). The grid is your operating model, meaning how the work actually flows, who owns each step, what finished means, and what good looks like inside your firm and nobody else's.

Artificial intelligence is the current, and a current does no work at all until it is wired to something designed to receive it.

A vendor who sells you wiring and calls it electricity has understood one step of the problem and mistaken it for the whole. That is simply a vendor selling the only part they can deliver, and that arrives with an invoice attached.

What the Demo Is Actually Showing You

I have been on both ends of this pitch. I have sat in the buyer's chair while a very good salesperson walked me through a very good demo, and I have spent years building the operating models these platforms claim to hand a firm inside a quarter, which is the vantage point that makes the demo difficult to watch.

The demo works and the plan makes sense because the demo firm is clean. Every process inside it has one owner, one path, and one definition of done. The platform did not create that order; it inherited it. What you are watching is software running on top of an organization that somebody built by hand, over years, before the software arrived.

Your firm is not that firm. The demo does not show the difference between the two, and the difference is why most AI projects have produced so little.

MIT's Project NANDA put a number on that in a report called “The GenAI Divide: State of AI in Business 2025,” which found that 95% of generative AI pilots produced no measurable impact on the profit line, despite $30–$40 billion in enterprise investment. The researchers put the cause on integration rather than model quality, on the failure to fit the tools into the way the companies actually worked.

What That Sale Buys an Uncodified Advisory Firm

Take a firm that never decided who owns the nonstandard client exception. You know the one. It gets handled a little differently in every office, by whoever happens to be closest to it, and nobody has ever written down which way is right.

Unify that firm's data, wire in and turn on the automation, and you have not resolved the inconsistency. You have industrialized it. The inconsistency now runs at speed, in every office at once, and you paid a seven-figure sum for the privilege.

The identical platform, dropped on a firm that did the unglamorous work first, is a genuine multiplier. Dropped on one that did not, the platform accelerates math that was already wrong, and the architecture debt the firm never paid down begins compounding at software speed.

Our own industry's numbers say the same thing in a quieter register. Schwab Advisor Services reported in January of this year that among the RIAs it surveyed, adoption of AI tools has more than doubled since 2023, with 63% now using them in some capacity. It also found that among advisors already using AI, only about 1 in 10 reported fully integrating it into their business strategy, while 82% of those users reported relying on generative tools, most often through individual experimentation rather than firm-wide systems.

Many firms have bought the tools. Very few have built the firm those tools were supposed to multiply. That is less an adoption problem than one of sequencing, and it is expensive in a way that surfaces late.

There is a second invoice coming for the unmapped firm that will arrive at an address many of the leaders buying the tech are not thinking about yet. The custom API platform you are proudest of, the custom integrations you lead the pitch with, the thing in your Frankenstack that you are certain is your edge, is probably the first line a buyer will mark down in diligence.

They open the hood, find the same stitched-together stack with someone copying and pasting between silos, and price the rebuild they are about to inherit. What reads as a strength from the seller's chair reads as a liability from the buyer's, and an unintegrated foundation is marked as a discount.

Why the Pull Is So Strong

I understand the temptation, and it is a rational response to an unfair choice. The wiring is the half you can buy; the other half is what your clients, your people, and your investors are actually asking about, and it can't be bought at any price.

Codifying how the firm runs, making the hard calls about who owns what, putting one person on an outcome and holding them to it. . . . All of that work is slow, political, frequently boring, and almost entirely invisible. Nobody hands a chief executive a trophy for finishing it; it carries no go-live date, and it will not fit on a board slide.

A platform, by contrast, arrives with an invoice you can show the board and a date you can circle on a calendar. It automates things, saves some time, and converts an unbounded organizational problem into a bounded procurement decision, and procurement decisions are the kind a leadership team knows how to justify.

That is the real product being sold. It lets a leadership team feel it has done the work without doing it, and that feeling is worth a great deal of money to the people selling it as well as the people buying it. The firm is not AI-ready 18 months later, but it is marginally more efficient at whatever it already was.

One Question to Ask the Vendor

The salesperson will be in your office soon, and the demo will be beautiful. When it is over, ask for the vendor’s process map. Not the architecture diagram and not the integration list, but the map of how work actually flows through the firm in that demo, with definitions at every step and concrete criteria for good at every handoff.

One of two things happens. Either the map appears, in which case you are looking at the actual product and can ask yourself honestly whether your firm has one of those. Or it does not appear, in which case what is being sold to you is order the vendor did not create, cannot transfer, and assumed you already possess.

The question costs you nothing, takes 30 seconds, and will end more unproductive sales calls than anything else you could say.

The Order to Spend In

When adopting AI, the order matters more than the purchase. A unified data layer is both necessary and nowhere near sufficient. The data lake is the wiring, not the current. Buy it once you have somewhere for the power to go, once the firm is codified, once one person owns each result and the definition of finished is written down rather than assumed.

Do it in that order, and the right platform becomes the multiplier it was sold as. Buy it first, and you have wired a building to a dark grid.

The current AI platforms are real, and I am not attempting to talk anybody out of using them. I am trying to keep you from mistaking the multiplier for the thing it multiplies because wiring with nothing plugged into it does not make a firm modern.

It just makes the dark more expensive.

Sean Baenen is co-founder of SMART Growth Partners. He was previously with NorthRock, which grew both assets and revenue at a 30% CAGR during his nearly six-year tenure as President and, later, leader of all M&A. He writes Built to Compound, a newsletter on growth and capital strategy for advisory firms, as well as House on Fire which offers leaders of growing firms principles for holding the line through the chaos.

A message from Advisor Perspectives and VettaFi: Discover something new! Click here to register for our upcoming webcasts.