Automatic support: Generate intelligent responses to people’s support questions. For example, in a Support category, the agent can use community content for context. A bonus is to have the answers generated through whispers so a responder can first use the suggested answers before replying
Category expert: Have a designated agent expert for each category in your community. Respond to the first post of a topic, with human follow-up if needed
Subtle moderation: Nudging users to post community-appropriate content in the right places. For example, if a topic seems like a mismatch with the category, instead of making that change automatically, replying in the topic suggesting other categories to alert the poster
Stalled topic follow-up: Automatically respond to topics that have gone unanswered for a period of time, using the stalled_topic trigger to nudge conversations forward
Due to the Agents being fully customizable, a community can leverage this feature in many other ways.
How it works
AI triage using Agent can unconditionally respond to topics and posts by using agents to understand context and drive intelligent responses.
To set up this feature, navigate to Admin → Plugins → Automations → New automation and select the LLM Agent Triage script. You must have an agent created before using this feature.
The following is an example automation rule and an associated agent.
Post created/edited: Fires when a new post is created or an existing post is edited
Stalled topic: Fires when a topic has had no activity for a configured period of time
Configuration
To configure this feature, you need to specify certain parameters:
Agent: AI Agent to use for triage. Must have default LLM and user set
Reply as Whisper: Whether the agent’s response should appear as a whisper
Silent Mode: If selected the agent will receive the content but will not post anything in the forum. An admin must create Custom tools before using this option. This is used in conjunction with the Tools field in the Agent settings. This can be useful for:
Making an HTTP request to a company internal API informing that something just happened
Performing side effects on Discourse such as creating a chat message, or flagging a post
Silent Mode and Reply as Whisper are mutually exclusive — only one can be active at a time.
Caveats
LLM calls can be expensive. Be careful to monitor costs and always consider running on small subsets
Be mindful that LLMs can hallucinate, so make sure to check any output carefully
Posts made by bots are automatically skipped and will not trigger the agent
Rate limiting is applied: by default, a maximum of 60 triage calls per minute globally and 2 per individual post per minute. These can be adjusted via the hidden site settings ai_automation_max_triage_per_minute and ai_automation_max_triage_per_post_per_minute
Do you have to exclusively use silent mode? Can I have reply as whisper selected? Im assuming it doesn’t matter as there is no reply (from a UI perspective we should probably disable certain options if it doesn’t depend on it)
Does it need to be only Forced tools? What about just regular tool calls from the persona? What if the persona has no Forced tools?
I think we should put a disclaimer for this option given the dependency on Custom tools and having forced tools selected via the Persona
Going back to the use cases I think this changes the name of this feature from AI auto responder back to AI triage via Persona since this option doesn’t have “auto-responses” that are visible or invisible (whispers)
Can something like “Writes” or “Replies” be added to the post page when the persona responds? Can something like that appear when normal users reply to the post be added?
Great feature, suggestion. In the first post maybe reference that the starting point is Discourse admin panel -> Plugins -> Automations -> Add automation and that the Automation feature is disabled by default and needs to be enabled.
Spent about 30 minutes looking through all the AI features to see how to enable this before noticing that this is under Automations (and I had to enable the Automation feature which was disabled by default).
Please note that this feature requires both the discourse-automation and discourse-ai plugins to function.
