Automating the week: a starting map

Ask most professionals where their week goes and they will describe their meetings and their main projects. Track it for a week and a different picture appears. A significant share of time disappears into work that sits between the meetings and around the projects: writing up, chasing, reformatting, collating, reminding, searching, copying information from one place to another.

This connective work matters. Without it nothing moves. But much of it follows rules, repeats every week and needs little judgement. That combination is what makes it a candidate for automation — whether through AI tools, the automation features already built into everyday software, or simply a better template and a fixed routine.

This playbook is a starting map. It does not try to cover everything that can be automated. It identifies the four categories of work that are almost always worth automating first, the two that never are, and a simple method for getting started without creating new problems.

Capture is the work of turning what happened into a usable record: meeting notes, action lists, call summaries, the record of a decision made in a corridor conversation. It is high-volume, it happens every day, and doing it by hand pulls attention away from the conversation itself.

It is a good first candidate because the quality bar is clear and easy to check. Either the record contains the decisions and actions accurately, or it does not. A person who was in the meeting can verify a summary in a minute.

Collation is the work of gathering information from several places into one: the weekly status report built from five project updates, the monthly pack assembled from three systems, the briefing note that pulls together what is known about a client before a meeting.

The labour in collation is almost entirely mechanical. The value lies in the result, not in the copying. Automating it frees time and, just as importantly, makes the output more consistent from week to week.

A surprising amount of professional time goes into moving things along: chasing an approval, reminding someone of a deadline, forwarding a request to the right person, checking whether an action has been done. This is the work most people dislike most, and the work most easily handled by simple rules.

Routing and reminders are often automated without any AI at all. Shared task lists with due dates, approval workflows, scheduled reminders and standard request forms already exist in most organisations' software. The barrier is usually setting them up, not the technology.

The fourth category is the starting point for recurring written work: the standard update, the routine reply, the meeting invitation with its agenda, the first outline of a familiar document. Starting from a blank page is slow; starting from a well-structured first version is fast.

This is where AI tools are most useful, and also where most care is needed. The first version is a starting point, never the finished item. The person who sends it is accountable for every word.

Some work should not be automated, however capable the tools become. The reason is not that it cannot be done by a machine. It is that the value of the work lies in a person doing it.

Decisions about hiring, performance, promotion, reward and exit shape people's lives and careers. Those affected are entitled to know that a person weighed the evidence, considered the context and is prepared to explain and stand behind the result.

Tools can help organise information for these decisions. They cannot carry the accountability, and they should not be allowed to appear to. When judgement about people is automated, trust in every other decision suffers too.

The second category is the work that builds and maintains relationships: the thank-you that is specific, the apology that is meant, the difficult conversation, the condolence, the promise made to a customer or colleague in your name.

People can tell when these are produced at scale, and the moment they can tell, the message stops doing its job. The point of a personal message is that a person took the time to write it. Automating it removes the only thing that made it valuable.

The categories above tell you where to look. The method below tells you how to begin without creating a new set of problems.

Three mistakes undo most early automation efforts.

The goal of automating the week is not to fill the freed time with more of the same work. It is to move time from the connective work to the work that needs a person: thinking, deciding, building relationships, developing others and doing the craft well.

Start small, keep the checks, protect the work that should stay human, and review honestly. A few well-chosen changes, made carefully, will return more time than an ambitious programme that nobody trusts.