Several people in our internal company chat today complained about Zammads automatic quoting feature: when you select text in an article and hit reply, Zammad will automatically quote this selected text in your reply.
This led me to realize that I also only trigger this feature accidentally. It usually happens when I copy-paste the name of the customer for the salutation (though we do have text modules with #{ticket.customer.lastname}, but that doesn’t work for all situations). Zammad will then automatically insert the selected customer name into my reply, which of course is not what I want. I usually then revert my changed, undo the selection, and hit reply again, but that is pretty annoying because it’s happening pretty often.
Maybe it would be more useful if the quoting feature had an explicit button to insert the quote (ie. floating atop of the selected text), instead of automatically triggering for selected text?
As an opposite example: Our team finds the feature extremely useful and clever as it enables you to quickly quote only relevant parts of the message instead of the whole one.
So personally, I would not like a button. But maybe one can provide a config option?
I would argue that the proposed solution doesn’t this make any harder. The way I imagine it would be exactly how it works here - double click the text and then the button right above it:
We had this same problem within our company as well. Our agents do not like this feature. I made a PR on Github, that adds an option to disable the feature a few days ago. This way, the feature is selectable. I hope it makes the next release.
I’d like to point out as well, that automatically quoting selected texts poses a risk for potential data leakage. screengrab|690x496
we’re currently discussing aboved mentioned PR and are wondering if this option rather workarounds other issues.
From my understanding the biggest issue with this function is that you can mark things outside articles which then is quoted by hitting “reply”. Did I oversee other use cases / show stoppers that create the need to disable such a standard behavior like in Thunderbird and Apple Mail entirely?
We’d like to understand the scopes better, when I tried to sum it up I kinda failed. :x
Marking things outside of articles is quite certainly an issue, but overall this feature is sub-optimal for our use case.
Our agents generally need to copy IDs or names and similar things from customer tickets to look things up in our systems, so they select to copy the information, work on the ticket itself, then hit reply and have the highlighted ID/name left as the single quote which is quite undesirable.
We too would like to see a possibility to disable this feature, I even personally went ahead and created a small custom package to add a reply button which automatically deselects any text to prevent this from happening.
Okay can we get away of issue descriptions like “annoying”?
We always aim at 80% user base so some functionalities or behavior may not change.
I’m still trying to find at which particular point the marking annoys.
Does it in general annoy all over the place or just out side of articles?
I still didn’t understand that specific point and that’s actually important to know to us.
Sorry, I didn’t mean to sound overly negative there, I will try to word it better next time.
Yes, generally all over the place, given the nature of our tickets we receive.
Selecting IDs or names from inside an article is often required to look up the user in question in our separate system for their support inquiry. Afterwards, one needs to make sure to not have selected anything before hitting the reply button; otherwise one has to manually remove the quoted ID/name from the reply if one forgot to do so.
The suggested mark-and-quote button like how it works here seems like a very good solution to this.
I’ll summarize the situations in which we had issues with the current behavior:
Copy-pasting the name of the customer from his original mail into the reply for the salutation.
(This is what I’ve mentioned in my OP, but I’ll describe it here in more detail.)
E.g. the customer signed his mail like this:
Yours sincerely,
Foo Bar
One might now double-click Bar and hit Ctrl+C to copy the name, so one can then paste it into the reply. As we provide several text modules for salutations to our agents, copy-pasting the name mostly wouldn’t actually be necessary, but:
not all agents might be aware of the text module feature
even if aware of text modules, copy-pasting the name might be ingrained in muscle memory
some of our customers send mails from generic accounts like Foobar Support <foobar.support@some-domain.com>. The text module is useless in this case because it thinks that customer’s last name is Support. Usually no agent will bother to edit the respective customer and fix the name, and everyone just copy-pastes the name from the last customer mail manually. Also in situations where multiple customer employees share a generic account, editing the customer wouldn’t really be viable.
Using the text selection absentmindedly while reading a mail, possibly to keep track of how far one has read a mail.
Copy-pasting paragraphs of a mail into an internal company chat, e.g. to get feedback from a colleague.
Copy-pasting e.g. a prior internal note of the ticket.
This is the one you already mentioned:
Curiously enough, I wasn’t even really aware that this is possible. I searched through our internal company chat and did encounter at least one complaint about exactly this behavior, where an agent had select an internal note and then wondered why it would show up as a quote in his reply.
A significant part of the problem is that agents often don’t realize what’s happening; in the original discussion in our internal company chat on 2020-05-19 (actually the discussion that led me to create this topic here) a colleague complained (loosely translated by me):
Please fix that replys sometimes only quote single words or sentences. It only works properly on the second attempt.
This probably makes this a rather significant UX problem, because from the POV of an agent who doesn’t realize that his text selection caused the quotes, this’ll make it look as if Zammad was buggy. This might unjustifiedly reduce the agent’s trust in the system (“if it’s mixing up my replies, what else might it mix up?”).
