A simple test for deciding which parts of your job are safe to speed up and which parts you should still do yourself.
The question I get asked most in workshops isn't "how does this work." It's "what am I supposed to use it for?"
People know the tools exist. They just can't see where they fit into a real week. So here's the test I use.
The two questions
For any task, ask:
- If this were 80% right, would that be useful?
- Would I know if it were wrong?
If the answer to both is yes, hand it over. If the answer to either is no, keep it.
That's it. It sounds almost too simple, but it sorts most of your work in about ten minutes.
Things that pass the test
First drafts of anything. A rough draft you can react to beats a blank page every time. You'll rewrite half of it. That's fine, rewriting is faster than starting.
Summarising things you'd otherwise skim. Long reports, meeting transcripts, policy documents, competitor sites. You'd have skimmed and missed things anyway. Now you get the shape of it and know where to read properly.
Reformatting the same message for different people. The board version, the team version, the client version. Same substance, different weight and length. This is genuinely tedious work and it disappears.
Structure and thinking-out-loud. Give it a messy brain dump and ask it to organise it into themes. You keep the thinking, it does the sorting.
Being your second pair of eyes. "Read this proposal as a sceptical finance director and tell me what you'd push back on." Not a substitute for a real colleague, but useful at 11pm when there isn't one.
Things that fail it
Anything where you'd be the last check. Legal wording, medical detail, financial figures, regulatory language, anything you'd sign. If you can't personally verify it, you can't use it.
Decisions that need to be yours. Who to hire. Who to let go. Whether to take the contract. You can use it to lay out the trade-offs, and that's genuinely helpful. Don't let it hold the pen.
Anything with confidential information in it, unless you know exactly where that data goes. This is the one people get wrong most often. Client names, staff records, unpublished figures, anything covered by a contract. Check your organisation's policy before, not after.
Relationships. The apology, the difficult feedback, the condolence note. People can tell. Use it to think through what to say if you're stuck. Then write it yourself.
The bit nobody tells you
When you first hand something over, it will take longer than doing it yourself. Briefing properly takes effort. You'll go back and forth. You'll think, I could have written this by now.
That's the cost of the first three times. By the fourth, you have a brief that works and it takes ninety seconds. The people getting real value out of these tools aren't smarter. They just pushed through the awkward stage on a handful of tasks they do repeatedly.
Pick two tasks. Not twenty. Two things you do every single week, that you find boring, that involve words. Get those two genuinely good. Then add another.
Where this leaves you
There's a version of this conversation that ends with "so your job is going away." I don't think that's the honest read.
What I see happening is more ordinary and more immediate. The admin layer around skilled work is thinning out. The judgement underneath it is worth more, not less, because there's suddenly a lot more mediocre output in the world and someone has to tell the difference.
You've spent years building the ability to tell the difference. Use the tools for everything underneath that, and keep the judgement where it belongs.
Where next
If this was useful, the live sessions are the same thing with your own work in front of you and someone to ask when it goes sideways.