I heard about this technique the other day. It’s an interesting way to do what sounds similar to creating an embedded command. I can’t say whether it would work in writing or not but it seems worth testing.
Of course I’d mark out the intended message the same way you’d do with embedded commands… very discretely.
Rather than try to explain it, I’ll let Steve G Jones who I heard about it from do so:
NLP Analog (Analogue) Marking Explained