๐๐จ๐ฐ ๐ญ๐จ ๐๐๐๐ง๐ญ๐ข๐๐ฒ ๐๐๐๐ฅ ๐๐ ๐๐ฉ๐ฉ๐จ๐ซ๐ญ๐ฎ๐ง๐ข๐ญ๐ข๐๐ฌ ๐ข๐ง ๐ญ๐ก๐ ๐๐ฎ๐ฌ๐ข๐ง๐๐ฌ๐ฌ (๐๐ข๐ญ๐ก๐จ๐ฎ๐ญ ๐๐ฌ๐ค๐ข๐ง๐ ๐ญ๐ก๐ ๐๐ซ๐จ๐ง๐ ๐๐ฎ๐๐ฌ๐ญ๐ข๐จ๐ง)
Most AI conversations go nowhere because they start in the wrong place. This post breaks down why asking about โrepetitive workโ shuts people downโand why focusing on friction gets to the real opportunities faster. It walks through how to spot where work actually breaks: delays, rework, missing information, and handoffs. Thatโs where AI starts to make a difference. If you want better outcomes, change the question.
Dawn Thiart
4/12/20261 min read


One of the mistakes I see in AI conversations is how they start.
โ๐ช๐ต๐ฎ๐โ๐ ๐ฟ๐ฒ๐ฝ๐ฒ๐๐ถ๐๐ถ๐๐ฒ ๐ถ๐ป ๐๐ผ๐๐ฟ ๐ฟ๐ผ๐น๐ฒ?โ
Sounds logical. Itโs not. If youโre opening with that, youโre off track.
๐๐ฒ๐ฟ๐ฒโ๐ ๐๐ต๐ฎ๐ ๐ต๐ฎ๐ฝ๐ฝ๐ฒ๐ป๐ ๐๐ต๐ฒ๐ป ๐๐ผ๐ ๐ฎ๐๐ธ ๐๐ต๐ฎ๐ ๐พ๐๐ฒ๐๐๐ถ๐ผ๐ป.
No one is going to sit there and tell you their job is repetitive โ even if parts of it are.
What they hear is:
โWhich part of your job isnโt valuable?โ
So the conversation stalls, you get safe answers and you miss the real issues.
๐ฌ๐ผ๐โ๐ฟ๐ฒ ๐ฎ๐๐ธ๐ถ๐ป๐ด ๐๐ต๐ฒ ๐๐ฟ๐ผ๐ป๐ด ๐๐ต๐ถ๐ป๐ด
Youโre not looking for repetition, youโre looking for friction.
โข where things slow down
โข where work gets stuck
โข where people have to redo things
โข where theyโre waiting on someone else
The problem isnโt the task, itโs everything around it.
๐ช๐ต๐ฎ๐ ๐๐ผ ๐ฎ๐๐ธ ๐ถ๐ป๐๐๐ฒ๐ฎ๐ฑ
Change the conversation. Ask things like:
โข Where do things tend to slow down?
โข What takes longer than it should?
โข Where do you have to go back and fix things?
โข What do you need before you can move forward?
โข Where do you spend time chasing information?
โข Whatโs the most frustrating part of this process?
This is where the conversation opens up.
๐ช๐ต๐ฎ๐ ๐๐ผ๐โ๐น๐น ๐๐๐ฎ๐ฟ๐ ๐๐ผ ๐ต๐ฒ๐ฎ๐ฟ
โข โIt dependsโฆโ
โข โWe have to check withโฆโ
โข โThe data isnโt always rightโฆโ
โข โWe end up redoing thisโฆโ
โข โWe wait forโฆโ
Thatโs the signal, thatโs where the real opportunity sits.
๐ง๐ต๐ถ๐ ๐ถ๐ ๐๐ต๐ฒ๐ฟ๐ฒ ๐๐ ๐ฎ๐ฐ๐๐๐ฎ๐น๐น๐ ๐๐ผ๐ฟ๐ธ๐
Not in the core task, but in the gaps.
โข Friction โ automate or orchestrate
โข Rework โ validate
โข Delays โ trigger earlier
โข Missing info โ assist
โข Inconsistency โ standardise
Thatโs where you get the real value. You can start to understand where to:
โข remove friction
โข improve flow
โข make work easier to get done
๐๐ถ๐ป๐ฎ๐น ๐๐ต๐ผ๐๐ด๐ต๐
If your AI conversations feel flat, itโs probably not the tech. Itโs the question.
Donโt start with tasks, start with friction.
Thatโs where the value is.