What do you still do manually, even though AI could do it?
AI can draft a reply, build a small app, or run a routine task. But being able to automate something does not always mean you want to hand it over.
Sometimes doing it yourself is how you notice a problem or understand what someone actually needs.
A few examples:
Replying to a frustrated customer.
Checking an app before sharing it.
Deciding what to build next.
I am building Inquir Compute for apps, databases and background jobs, so I am curious about where people draw that line.
What is one task you deliberately keep manual, and what would need to change before you trusted AI with it?
155 views
Replies
If something is remotely close to being creative, I do it myself.
I am technical but my code is almost 100 percent ai, my internal docs are 90 percent ai - I always do proofreading and making the doc language better- I do posts, comments emails anything external facing manually
Does rewriting an AI draft in your own voice take longer than starting from a blank page?
I wonder if the useful split is letting it check a finished post for missing context while keeping the first draft yours
@al_kub Mostly not in my case. I am always careful and I try to do a writing standard for my ai generated docs. For example I always use asd-ste100 type voice in my technical docs.
Honestly, less and less. Code, translations into 50 languages, store copy, SEO pages: almost all of it goes through AI now. What I keep for myself is the last step. I listen to the songs our app generates before I trust a change to the prompts, and I press "publish" myself on anything that goes out under my name. AI could do both. But "does this song actually feel good?" is the one judgment I can't check by reading a report, and it's the whole product.
@gbasovv Do you compare the old and new prompts on the same set of requests before changing them?
I would want a few awkward examples in that set too so one good song does not hide where the change made things worse
picking the app icon out of a batch of options, something about it just has to feel right in a way i cant put into a prompt
Reading what changed in a tool update before I let the agent use it. I tried to hand that off and couldn't, because what changes is a sentence of description and only a person can say whether the new sentence is fine. So I automated the stop, not the decision. It refuses the call until I've read it.