Automation
5 ways automation creates capacity for client work

The business case for automation is usually written in headcount. That is the wrong unit. In most service businesses the scarce resource is not people. It is the attention of your best people, and how much of it reaches a client. Automation earns its keep where it moves attention from internal work to client work. Here are five places that reliably happens.
1. Stop entering the same thing twice
The most common automation in any business is also the least glamorous: taking a piece of information that has been typed once and stopping it from being typed again. A client’s details go into the CRM, then into the engagement letter, then into the finance system. Each retype costs a few minutes and introduces a small chance of an error that somebody spends an hour finding later.
The fix is rarely a platform. It is usually a connection between two systems you already pay for, plus a decision about which one is allowed to be wrong.
2. Let approvals chase themselves
Approvals tend to be fast in the doing and slow in the waiting. A partner signs off in ninety seconds; the request sat in an inbox for four days first.
Automating the chase usually recovers more calendar time than automating the decision ever would: a reminder that goes out on its own, a visible queue, an escalation after a set period. It also removes the least pleasant task in the business, which is a junior person having to nudge a senior one repeatedly.
3. Assemble the report, do not rebuild it
Most recurring reports are not analysis. They are assembly: pulling the same figures from the same three places into the same layout, then writing two paragraphs of commentary at the end.
The assembly is mechanical and should be treated that way. The commentary is the part that needed a person, and it is usually written in the last twenty minutes by someone who has spent four hours copying cells. Automating the first part does not remove the judgement. It gives the judgement more room.
4. Make onboarding a process rather than a memory
Client and staff onboarding tends to run on a checklist that lives in one person’s head, supplemented by a document that is slightly out of date. It works, right up until that person is on leave during a busy week.
Turning onboarding into something that triggers its own requests, accounts, introductions and follow-ups has an obvious efficiency argument. The stronger argument is consistency: every client gets the version of your firm that your best person would have delivered.
5. Automate the first draft, not the final answer
This is where AI belongs in an automation conversation, and where it is most often misapplied. A blank page is expensive. A first draft that is seventy per cent right, and clearly labelled as a draft, is cheap, and a skilled person takes it the rest of the way in a fraction of the time.
The line matters. Automate up to the point where judgement starts, and stop there. Anything going to a client should have had a person’s attention on it. The goal is that their attention arrives at the interesting part rather than the formatting.
Where to start
The instinct is to pick the biggest time sink. The better instinct is to pick the one your team complains about most. Automation that removes an irritation gets adopted; automation that removes a task nobody minded doing tends to get quietly worked around.
Beyond that, look for work that is repetitive, predictable and easy to check: the monthly report, the approval chase, the details being carried from one system into another.
Whatever you pick, measure the before: how long it takes today, how often it happens, and how much of that time would go if the process changed. The most common reason automation work cannot prove its value later is that nobody wrote down what the old way actually cost.
A small automation is easier to justify than a large one, because the result arrives quickly and in a form somebody can check. The aim was never to automate as much as possible. It is to have more of your team’s attention pointed at the work that needs them.
Keep reading
More insights.

AI & Data
When AI stops answering and starts doing
AI & Data
Teaching an assistant how your firm does something
AI & Data
A seat for everyone, or a system that does the work

Strategy
Where client information goes when you use AI