I'm actually totally stumped by the whole thing. OCR doesn't even make sense, because OCR is terrible at handwriting generally. With forms they usually require you to write block letters and numbers inside of a kind of separate grid for each field. And maybe fill in some bubbles too. Anything anywhere on the page outside of the form fields is ignored.
I'd find it far more plausible that they print all letters, those including forms and not, on the same template, and that returned forms get some kind of bar code or status stamped on the bottom upon being received, so they need to keep it empty for that. Kind of like how US envelopes get a little bar code printed on them by post office sorting. I have no earthly idea whether that's closer to the real reason though.
If you look up some of these letters you'll see they have the quasi-official-looking things you'd otherwise see on scam letters, like a stamp that says "Enforcement Visit Approved" with a signature on it.
I think "do not write below this line" is just another one of those things, it makes the letter seem like its part of Official Serious Bureaucracy.
I think there's a pretty reasonable explanation here, which is that "Do not write below the line" is a genuine instruction, but not for the recipient of the letter.
Post offices may make notes such as "undeliverable" on a piece of mail. The sending company may make changes to their mailers which must be hand-updated on pre-printed cards. In both cases, writing below the line may obscure which ID had its letter rejected by the Post or which IDs have not had updated mailers sent yet.
By the time the recipient of the letter receives it, they may write below the line as much as they like, as the instructions have already been followed by those they were intended for.
I would not expect first line support to be aware of this.
That is true, but OCR is nonetheless used in many situations like this, for example at the postal office (the US postal office started doing this in 1965). Even if they can recognize only a fraction of the letters, it is a huge savings in terms of processing costs. The remaining will be handled manually anyway.
I'm actually totally stumped by the whole thing. OCR doesn't even make sense, because OCR is terrible at handwriting generally. With forms they usually require you to write block letters and numbers inside of a kind of separate grid for each field. And maybe fill in some bubbles too. Anything anywhere on the page outside of the form fields is ignored.
I'd find it far more plausible that they print all letters, those including forms and not, on the same template, and that returned forms get some kind of bar code or status stamped on the bottom upon being received, so they need to keep it empty for that. Kind of like how US envelopes get a little bar code printed on them by post office sorting. I have no earthly idea whether that's closer to the real reason though.