According to support, the Writing Assistant currently only uses the selected text as its source of information. Internal notes are not taken into account.
For our support workflows, it would be very helpful if the Writing Assistant could optionally use internal ticket notes as additional context. In many cases, colleagues already document their analysis or the final solution in an internal note.
A particularly useful feature would be the ability to select text directly within an internal note and use it as the basis for the Writing Assistant. The assistant could then generate an appropriate customer-facing response from that information using a prompt, without requiring agents to manually copy and rephrase the content.
Is such a feature already planned or on the roadmap?
I think a writing assistant is currently only designed as a text tool that should help with writing. Not really about using content from other places.
But in general, we will maybe think about having a different way for such a situation in the future.
Yes, that’s how I understand it as well, and I already find the current functionality very useful.
Precisely for that reason, I believe this would be a valuable next step. At the moment, the agent still has to manually write the response, even though the relevant information is often already documented in internal notes or elsewhere within the ticket.
A significant improvement would be the ability for the writing assistant to optionally use internal notes as additional context. Even better would be the option to select text directly from an internal note and have the assistant generate a customer-facing response based on that information and the provided prompt.
This would save time, improve consistency in customer communication, and make better use of solutions that have already been documented by colleagues.
Are there any plans or discussions to include such functionality in a future release?
I agree current functionality already very useful and also agree increasing sources of different text etc fields that can be input for LLM processing sessions would offer more value.
Besides using internal notes as additional context I cannot find how to use the customers and previous sender’s email body message content as additional context or the customer data fields ie /#ticket/zoom/12345
In version 7.1 the writing assistant at /#ai/text_tools already features " Writing Assistant Tool" with fields for “Custom instructions” (Provide specific and unambiguous instructions for the LLM to process a given text which will be part of the system prompt.)
Engineering the “Custom instructions” field to support Merge Tags / Form Fields / Template Variables / Named Arguments / aka Placeholders representing the Zammad data ie {{note.body,internal}} or {{message.email}} or {{ticket.articles.content}} which sends to LLM actual additional context allows a reasonably simple feature iteration release without significant code development.
The above solution would be beneficial for Zammad users to quickly experiment with new creative writing assistant solutions
Plus, Zammad developers don’t need to worry and think about different situations or different ways.
