AI took over this part of my job. Here's what I do now.
An anonymous account from a reader who used to spend half their week on work that's now mostly automated.
I used to write first drafts of client reports. That was maybe 40 percent of my week, gathering the data, structuring it, writing it up in plain language. Now a tool does the first draft in about ninety seconds, and it's genuinely fine, not perfect, but fine enough that rewriting it from scratch would be a waste of everyone's time.
The first few months after that shifted, I didn't have a plan. I just did the same job with a new tool bolted on, and it felt precarious, like I was one product update away from being unnecessary. That feeling turned out to be the useful part. It forced me to actually look at what my job was for, not just what my tasks had been.
What I landed on: clients weren't paying for the drafting. They were paying for someone to know which numbers mattered, catch the ones that looked wrong, and translate the whole thing into a recommendation they could act on. The drafting was just the visible part of that. Once the visible part got automated, the actual value, judgment, context, catching mistakes, didn't go anywhere. It just became a bigger share of what I was doing relative to everything else.
What actually changed day to day
I spend less time typing and more time reviewing. Less time structuring and more time deciding what the structure should emphasize for a specific client. I read the AI draft the way an editor reads a junior writer's draft, assuming it's competent but not assuming it's right.
I also started doing more of the client-facing conversation that used to get squeezed out when drafting took up so much of my week. That part, explaining the recommendation, answering the "why does this matter" question, turned out to be the part clients actually remembered.
What I'd tell someone this just happened to
Don't treat the automated task as your identity. It was never really the job, it was the visible mechanism of the job. The uncomfortable version of this is realizing how much of your week was mechanism versus judgment, but that's useful information, not bad news. It tells you exactly where to spend the time you got back.
And don't assume the shift is finished. Mine wasn't a one-time event, it's kept moving, a little more each year. Staying ahead of it means treating your task mix as something to actively manage, not something that happens to you.
See how your own task mix scores
Frequently asked questions
Did this reduce headcount on your team?
Not immediately, though it did change how new hires get trained, less time learning to draft from scratch, more time learning what makes a draft actually right for a client.
How long did it take to feel stable again?
Roughly six months from "this feels precarious" to "this feels like a normal part of the job." The instability was mostly about not having reframed the role yet, not about the technology itself.
← Back to the assessment · Related: How to reskill for an AI economy without quitting your job