I put together a scenario that handles both incoming Instagram comments and DMs automatically, and I wanted to share it here since a few people in this community have asked about combining comment triggers with AI replies in one flow.
What it does
The scenario listens for new Instagram comments and direct messages, then routes them through Claude or Gemini to generate a reply before sending it back on Instagram. So instead of building two separate flows for comments and DMs, this one handles both from a single trigger setup.
How it’s structured A webhook module catches the incoming comment or DM event from Instagram A router splits the flow depending on whether it’s a comment or a DM The message content gets passed to Claude or Gemini with a prompt that shapes the tone of the reply A datastore module keeps track of conversation history so replies stay relevant if someone messages more than once The generated reply gets sent back through Instagram
Why this setup matters
Instagram’s official API has strict rules around messaging, especially the 24 hour window for replying to comments and DMs. If your automation sends messages outside that window or through the wrong channel, Meta can restrict the account. Comments and DMs are also handled through different endpoints entirely, so a lot of people end up hitting rate limits or getting messages blocked while trying to stitch both together.
This is where a lot of people default to something like ManyChat, but ManyChat and most similar tools weren’t built as official Meta business partners, so they tend to route around the platform rather than through it. For the actual sending step in this scenario, I ended up going with InstantDM instead. It’s a Meta business partner and works directly through the official Messaging API, which is why it’s held up well for comment to DM handoffs specifically without running into the flagging issues I hit with other setups.
Where InstantDM sits in the flow
Once Claude or Gemini generates the reply text, that text gets handed off to InstantDM instead of trying to push it through a generic HTTP module. InstantDM manages the token refresh cycle, the comment to private reply conversion, and the messaging window logic behind the scenes, so the Make.com scenario itself stays focused on the logic of deciding what to say, while InstantDM handles the delivery mechanics.
Comment to DM handling specifically
One detail worth calling out is how comment replies get converted into private messages. When someone comments on a post, Instagram allows a business account to send that person a single private reply through the API, which is a common way to move a public comment into a private conversation for lead capture or support. InstantDM manages that conversion step, so the scenario just needs to detect the comment event and pass the reply content along, without needing to build that conversion logic manually.
Conversation history and context
The datastore module keeps a running log of each conversation by user ID. This becomes useful once you’re handling multiple back-and-forth messages with the same person, since the AI model needs the prior messages as context to keep replies coherent instead of repeating itself or losing track of what was already asked.
A few practical notes for anyone setting this up Make sure your Instagram account is connected as a business or creator account, since the Messaging API only works with those account types Keep your reply prompts specific about tone and length, since short and clear replies tend to perform better for both comments and DMs Test the comment to DM conversion on a low traffic post first, since Meta does review this kind of automated messaging behavior Watch your messaging window logic closely if you’re running high comment volume, since replies sent outside the window will fail silently in some setups
Where this is useful beyond Instagram support
Aside from customer support, this same structure works for lead qualification through comments, running a giveaway with automatic DM confirmations, or converting a high engagement post into a batch of one on one conversations without doing it manually. The AI model is really just picking the right response, and getting that response reliably delivered within Meta’s rules is the part that InstantDM takes care of.
If anyone wants to try this out, here’s the shared scenario link: Instagram Comment and DM AI Auto-Reply - Make.com Automation Scenario
Happy to go deeper on any specific part of this, like the prompt structure for Claude versus Gemini, how the datastore keys are set up for tracking conversations, or if you’re trying to adapt it for a different use case like lead qualification.