# About system and AI bot personal messages **URL:** \ **Category:** Support\ **Tags:** personal-messages\ **Created:** [October 9, 2026, 1:58pm UTC](https://meta.discourse.org/t/about-system-and-ai-bot-personal-messages/414347 "2026-10-09T13:58:31Z")\ **Posts on this page:** 2\ **Page:** 1 ## AI-generated summary The discussion addresses how to effectively manage and track personal messages (PMs) from system accounts and AI bots without resorting to console access or user impersonation. [satonotdead](/t/-/414347/1) questions whether Discourse workflows can be used to forward these messages, noting that while recent commits show promise, they may not fully solve the issue. They propose that adding an HTTP node to invite staff to these messages might be a solution, though they consider it potentially over-engineered. The goal is to find a simple, reliable method to track feedback from these automated entities without disabling their ability to message users, which would create inconsistency. [putty](/t/-/414347/2) shares their experience with a workflow that monitors PMs from specific agents. They suggest that it is easy to add a step to invite staff members to these messages. In their setup, they auto-tag these messages to easily identify them, providing a practical example of how workflows can be utilized for this purpose.
**Author:** ![satonotdead](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/satonotdead/32/447830_2.png) [@satonotdead](https://meta.discourse.org/u/satonotdead)\ **Post date:** [October 9, 2026, 1:58pm UTC](https://meta.discourse.org/t/about-system-and-ai-bot-personal-messages/414347/1 "2026-10-09T13:58:31Z")
> [@Discourse MiniMod](https://meta.discourse.org/t/discourse-minimod/400019/9): > > That’s actually a placeholder account. You should flag the topic next time. Coming from that reply and also a [recently created topic](https://meta.discourse.org/t/are-the-ai-agent-chats-working-here-on-meta/414345), I wonder how to handle the personal messages on ‘system’ and all the AI bots without the need to use the console and/or impersonate those users. Is workflows currently capable of forwarding them and is this the suggested approach? I saw new commits related (and I am very excited about them) but I think they were not actually designed for this use case and do not solve the issue I am bringing here to discuss. Perhaps also adding an HTTP node to invite staff members to those messages could work, but I feel that is a little over-engineered. > **What I want?** Track _system_ and bots personal messages in a simple, reliable and suggested approach. Disabling their mentioning and PMs could disable simple methods that users can follow to share feedback (something we want to be able to do) and adds inconsistency with the rest of the users as a side effect. Open to discuss it and read thoughts, creative ways or workarounds. ---
**Author:** ![putty](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/putty/32/370902_2.png) [@putty](https://meta.discourse.org/u/putty)\ **Post date:** [October 9, 2026, 4:42pm UTC](https://meta.discourse.org/t/about-system-and-ai-bot-personal-messages/414347/2 "2026-10-09T16:42:28Z")
> [@satonotdead](#): > > Perhaps also adding an HTTP node to invite staff members to those messages could work, but I feel that is a little over-engineered. I have a workflow that watches for PMs from a specific agent on our site. It would be easy to add a follow-up “add staff to this message” in that case. For my use, I auto-tag all these messages so I can easily identify them. Granted, it works because my agent auto-replies. It would get tricky for something that’s “watching” a system account that won’t auto-reply. ![CleanShot 2026-10-09 at 12.41.24](https://global.discourse-cdn.com/meta/original/4X/2/3/b/23be721d56c3a93865e5626a1d092f3dec070e41.png)