AI can help with the work you keep putting off
There’s a particular kind of work that keeps slipping down the list. It matters, but getting started means digging through emails, finding the right spreadsheet and pulling together a dozen little pieces.
So it waits. And next week, it’s still there.
I recently read Dharmesh Shah’s article, How I Use AI To Do Things I Don’t Do, But Should. What appealed to me was his idea that AI can expand our capacity: help us make progress on useful work we haven’t been finding time for.
That feels like a practical place to start.
Give the fiddly work a starting point
Think about the process you’ve been meaning to write down. Or the meeting notes that need turning into clear actions. Or the customer questions scattered across emails that could become a useful FAQ.
With the right material and instructions, AI can help organise those pieces and prepare a first draft. You then have something concrete to work through, rather than a blank page and a pile of information.
For example, you might ask your chosen AI:
Use these notes to draft a checklist for a new team member who hasn’t done this task before. Use plain English and put one clear action in each numbered step. Explain any unfamiliar terms. Flag anything unclear or missing rather than inventing steps.
The starting material matters. Shah’s linked article on context makes that point well. Give AI the relevant information, explain who the result is for and show it what a useful answer looks like.
Then put your own eye over it
This is the part I’d keep firmly in human hands.
Check the facts. Look for missing steps and assumptions. Make sure the result is consistent, concise and sounds like you. A tidy draft still needs to reflect how your business actually works.
My suggestion is to choose one task you keep putting off and try this approach. Use material you’re comfortable sharing with your chosen tool. Then judge the whole exercise, including your checking and editing: did it help you finish something useful?
If it becomes a reliable recurring process, you could consider scheduling it, as Shah explains in his article on scheduled AI tasks. Start with a reviewed draft before giving it more responsibility.
For me, that’s an appealing use of AI: making room for worthwhile work, while keeping our own judgement and voice in the result.
This post was inspired by Dharmesh Shah’s article, published on simple.ai on 2 October 2026. The practical examples and human-review approach here are my own take on that idea.