Repeat mistakes start where information gets copied
Cody Halovich · 16 November 2026 · 4 minute read

The owner of an advisory firm tells us the same person keeps making the same mistakes. An old address on a transfer form. A KYC update that reached the CRM but never the custodian. A client review prepared from last year's notes. Each one gets caught, corrected and talked about, and a month later another turns up that looks just like it.
A repeated mistake points at a step, not a person
When an error happens twice, it is tempting to treat it as a performance problem. Look at where it happened instead. Careless people make mistakes everywhere. Careful people make them in the same few places, and those places share one feature: a fact has to be read in one system and typed into another.
Opening one new account at an advisory practice shows how many of those places there are. The client's details go into the CRM. The same details go into the custodian's portal and onto the transfer form. The planning software needs them again. A tracking spreadsheet gets a row, and half the supporting documents arrive by email. Five systems, one client, and the same address entered three or four times by hand.
Every switch asks someone to hold a fact in their head
Copying between systems is not one task. It is find the record, read the value, switch screens, find the field, type it from memory, and confirm it went in. Somewhere in that sequence the phone rings, a colleague asks a question, or another client's email lands. The person comes back to the screen holding a value that may now belong to someone else.
That is not a character flaw. It is how people work when they move between files. In one experiment, practising auditors read inventory memos for two different companies and then attributed evidence to the wrong company 40 percent of the time (Lindberg and Maletta, summarized in Current Issues in Auditing, 2012). These were trained professionals doing work they knew well, and moving between two clients was enough. Your team moves between many clients and five systems every day. Each crossing carries the same risk of one client's facts landing in the next client's file.
A second check catches some of it, not all of it
The usual fix is a second pair of eyes. Someone reviews the transfer form before it goes out, or a checklist gets added to account opening. That helps, and it also adds a step. The reviewer is switching between files too, and an interrupted reviewer is a filter with holes in it. Errors still get through, now with two people's time attached.
Replacing the person fails the same way. The next hire sits down at the same five screens and copies the same address three times. Within a few months the same mistakes come back with a new name beside them. The step never changed, so the result was never going to change either.
Give each fact one home and let the systems move it
The fix that holds is to take the copying out. Pick one system as the home for each client fact: the address, the KYC details, the beneficiaries. When a fact changes, it changes there, once. An automated workflow carries it to the custodian, the planning software and the tracking sheet, and flags anything that does not match.
Nobody has to remember which screen they left. Your people keep the work that needs judgment: the KYC conversation itself, the unusual transfer, the client whose situation just changed. What leaves their desks is the retyping, which is where the mistakes were living.
This matters before you add AI to any of it. A tool dropped on top of the same five screens becomes a sixth place to copy from and one more thing to check. Give the facts a home first, then automate how they move.
Start with the last ten mistakes
This week, pull the last ten mistakes that reached a client or were caught at the last moment. For each one, write down the task it happened in, count the systems that task touches, and mark the exact step where it went wrong. Put a tick beside every step where someone was copying information from one screen into another.
Then look at which field shows up most often among the ticks. Trace every place that field gets typed by hand. That list is the first workflow to fix, and it will tell you more about the mistakes than another conversation with the person who made them.
The next step
Have a question this didn't answer?
The discovery call is free. We'll talk about your operation and tell you honestly whether AI automation is the right move.