How to Use ManyChat for Instagram DM Automation

Open ManyChat, connect an Instagram account, set a single keyword to auto-DM a link, and call the automation done. That's how most brands set up Instagram DM automation, and it's also why so many of these flows quietly stop working within a day: the DM goes out fine, but nobody qualifies the person, nobody routes them anywhere useful, and half the follow-ups never send because they land outside the window Meta allows for automated messages. ManyChat can genuinely run a comment-to-DM funnel, a keyword flow, and a story-reply sequence from one account, but the tool doesn't enforce good structure or policy compliance on its own, it just makes it easy to publish a flow that looks finished and isn't.

For background on what ManyChat is and what its plans include before building this workflow, see our full ManyChat profile. If you're setting this up for the first time, plan for an hour of setup rather than the "five-minute automation" pitch, because the parts that actually matter (routing, timing, and compliance) take longer than dragging a trigger onto a canvas.

Step 1: Set Up a Comment-to-DM Automation on a Specific Post, Not the Whole Account

In ManyChat, this lives under Automation, then Instagram, using the Post/Reel Comments trigger. Pick one live post or reel, attach a specific keyword to it, such as "GUIDE" or "PRICING," and configure two things: a public reply that posts under the comment itself, and a private DM that follows it. The public reply matters more than it looks like it should, because a visible response under a comment signals to other people scrolling that the account actually engages, which pulls in more comments on the same post.

The mistake is applying one comment-to-DM automation across every post going forward instead of building it per post, per campaign. A keyword that makes sense on a giveaway post reads as noise attached to an unrelated product photo two weeks later, and a generic "any comment" trigger set up account-wide means every unrelated comment, including complaints or spam, fires the same automated DM. Build the automation around one post's specific offer, and set a new one each time a new post needs its own trigger, rather than leaving a single rule running indefinitely in the background.

Step 2: Build a Keyword-Triggered Flow for Direct DMs and Story Replies

Comment triggers only cover people who comment publicly. A meaningful share of engagement on Instagram happens in DMs and story replies instead, and ManyChat's Keywords feature, along with the Story Mention Reply and Story Reply triggers, catches that traffic. Set a keyword so that anyone who sends that word directly to the account, replies to a story, or reacts to a story with an emoji, enters the same flow a comment would.

Build the flow itself in the Flow Builder as an actual sequence, not a single message block: a trigger, a first message that confirms what the person is getting, then a branch based on what they need next. A flow that's just one keyword firing one static reply is barely different from an auto-responder, and it wastes the part of ManyChat that's actually useful, which is branching a single entry point into different paths depending on how someone answers. Map the flow out before building it in the canvas: what the opener says, what the two or three most likely responses are, and where each of those responses needs to lead.

Step 3: Use Quick Reply Buttons to Qualify and Route People, Not Just Hand Out a Link

The most common failure in an Instagram DM automation is sending everyone who triggers it the exact same link regardless of what they actually want. Quick Reply buttons fix this cheaply: instead of one message with one link, the first DM offers two or three buttons ("I'm just browsing," "I want pricing," "I want to book a call"), and each button routes to a different next step in the flow. This does the qualifying work a human would otherwise have to do manually, and it means the eventual link, offer, or handoff to a real person actually matches what the contact said they wanted.

Use buttons for anything with more than one reasonable answer, and reserve free-text keyword capture for cases where you genuinely need an open response, like an email address or a specific product name. A flow full of open text fields is harder to route automatically and produces messier data than one built around a short set of predictable choices, and predictable choices are exactly what a button-based flow is built to handle.

Step 4: Respect Meta's 24-Hour Messaging Window Before It Breaks the Flow

This is the step most comment-to-DM setups skip, and it's the one that determines whether the rest of the flow ever gets seen. Meta only allows automated messages to a person within 24 hours of their last interaction with the account, whether that interaction is a comment, a DM, or a story reply. If the person replies within that window, the 24-hour clock resets and the conversation can continue; if they don't, ManyChat will not deliver further automated messages to them, and after that a 7-day window opens where only a manual message sent from the Inbox by a real person can reach them, not another automation.

Design the flow around this instead of around an ideal follow-up cadence. A flow that assumes a 3-day drip sequence will silently fail its later steps for anyone who doesn't reply within a day, and ManyChat will hold those messages back rather than sending them out of policy, which looks like the automation is broken when it's actually just complying with Meta's rules. Put the most important message, the one thing you actually need someone to see, inside that first 24-hour window rather than several steps downstream, and treat anything past it as something a real person follows up on manually, not automation.

Step 5: Add a Growth Tool as an Entry Point Beyond Comments and Keywords

Comment-to-DM and keyword triggers both depend on someone already engaging with a specific post or story. ManyChat's Growth Tools give the same flow other entry points that don't require that first engagement, most usefully a link-in-bio style widget that opens straight into the DM flow when someone taps it from the account's bio, an ad, or a landing page. This matters for traffic that never touches a comment section at all, someone clicking from a paid ad or a bio link goes straight into the qualifying flow instead of needing to comment a keyword first just to get started.

Point a Growth Tool at the same flow already built in Step 2 rather than creating a separate one, so a contact coming from a bio link and a contact coming from a comment end up qualified and routed the same way. Running two different flows for two different entry points into the same offer doubles the maintenance work for no real benefit once the underlying qualifying questions are the same.

Step 6: Review Open, Click, and Conversion Rates, Then Refine the Flow

A flow isn't finished once it's published, it's a hypothesis about what triggers, buttons, and messages will actually move someone from a comment to a customer. ManyChat's Analytics for each flow shows delivery, open, and click rates per message, and where in the sequence people are dropping off. Check this on a regular cadence, weekly while a flow is new and monthly once it's stable, rather than assuming a flow that worked at launch is still working two months later.

Use the drop-off data to fix the actual failure point instead of rewriting the whole flow. A low open rate on the first DM usually points at the comment trigger or the public reply not being compelling enough to pull people into the conversation at all; a good open rate with a low click rate points at the message copy or the button labels not matching what people expected when they triggered the flow; a good click rate with no downstream conversions points at the offer on the other end of the link, not the automation that got someone there. Fix whichever of those three is actually broken, then keep the parts that are working rather than replacing them along with everything else.

What This Workflow Actually Produces

Done properly, this isn't one keyword that fires a static DM to anyone who happens to comment, it's a set of specific triggers, comments, keywords, story replies, and a bio link, all feeding one qualifying flow that routes people by what they actually asked for, respects the messaging window Meta enforces, and gets checked against real open and click data on a regular cycle. That's a meaningfully different setup than a single automation left running indefinitely, and it's the difference between an Instagram DM funnel that keeps converting and one that quietly stops reaching anyone after the first day because nobody built it around the 24-hour rule in the first place.

If you're still deciding whether ManyChat is the right tool for this specifically, or how it fits alongside the rest of an Instagram content and ad strategy, our ManyChat tool profile covers the plans, pricing, and platform limits in more detail.