The Transformation List

The quickest early win is usually the reporting nobody reads twice: weekly updates, recap emails, spreadsheets rebuilt by hand every Friday.

In a poll of operations staff¹, respondents said nearly a third of their week went to compiling information other people had already written down somewhere. It topped the list of tasks they wanted to hand off.

At Northwind, regional managers once spent Monday mornings stitching together sales figures. An assistant now drafts the summary overnight, and the meeting opens with decisions instead of slides.

Assistants are only as good as the files they can read. Outdated wikis, duplicate folders and orphaned docs¹ lead to confident answers that are simply wrong, and trust erodes fast.

Teams that archived stale pages and labeled the trusted ones first got sharper replies from day one, and spent far less time arguing about which version was right.

Some tasks look like overhead but carry meaning, whether that is a handwritten note to a client or a careful review of a teammate's draft.

At Hooli, a support group let an assistant write first drafts but kept the final reply human. Ratings climbed, and the team felt the job was still theirs.

Anyone can now assemble a working tool in an afternoon. The risk is that nobody with deep experience looks at it before it reaches customers or touches the books.

Teams at Initech assign each internal build a reviewer from finance, legal or security who signs off before launch.

Assistants can lay out options quickly, but the reasons a team picked one over another tend to vanish once the meeting ends and people move on.

Groups that logged the reasoning behind major calls learned faster and onboarded new members more easily.

Most job descriptions were written before assistants arrived. Many no longer match how people actually spend their days, or what their managers expect.

Teams that rewrote them annually found it easier to hire, set goals and explain to staff how their roles were evolving.

Assistants make it tempting to add scope without adding people. Past a certain size, though, the cost of keeping everyone in sync swamps whatever time the tools give back.

The teams we studied did best when they stayed under nine members.

A single policy for every case ends up too strict for drafting emails and far too loose for anything that touches customer records.

Teams at Globex label each use by risk level, so low-stakes work needs no approval and sensitive work gets a quick review.

Requests to try a new tool often sit in a review queue for weeks, and in the meantime people quietly sign up for something else on their own card.

A firm deadline for each review keeps the process honest.

Official process maps leave out the side chats, private spreadsheets and small favors that actually keep work moving.

Teams that followed one real request end to end found the slowest step was often one nobody had listed.

Trials tend to run on tidy sample data, then break in the first week they meet real customers and real mess.

Feeding new workflows the oddest examples on file exposes weak spots early.

Knowing a particular tool matters less every year. Spotting a flawed answer and knowing how to fix it matters more.

Hiring teams at Vandelay now give candidates a draft with planted errors and ask them to find and explain each one.

Cheap output makes raw volume a poor signal. The colleagues worth promoting are the ones whose work lifts the whole team.

Rubrics that still count pages or tickets closed end up rewarding busywork.

Canned examples look great and fade by Friday. Sessions built on a team's own files and tickets stick.

Brief weekly practice outlasts a single long course.

Newcomers once learned by doing simple tasks that software now handles. Remove those and the first rung disappears.

Some groups have juniors review and correct machine drafts as a deliberate training step.

Open-ended trials eat attention for months. A fixed end date forces a clear choice to expand, rework or drop.

Groups that ended more trials ended up launching more of the good ones.

When only wins get airtime, people bury their misses and others repeat them.

A short, no-blame note about a failed trial often teaches more than a success story.

Adoption rises when a leader walks the team through a real memo or plan built with an assistant. It falls flat when the only signal is a mandate.

Showing the mistakes matters more than a perfect demo, because it tells others that trying is safe.

New routines usually feel slower before they feel faster. Groups that expect that slump stick with it, while others quit good ideas too soon.

Saying it up front, with a rough timeline, stops people from mistaking a normal dip for failure.