You post a reel of a new set of gel extensions on Thursday afternoon. By Friday morning there are 34 unread DMs. Half say "price?", a few say "how much for these", three ask whether you have anything on Saturday, one is a supplier, one is spam, and one is a regular asking whether her appointment is still on. The post did its job. Now someone has to sit and type replies, and every hour that passes, a few of those people lose interest.
Instagram DM automation means software reads and answers those messages for you. With Klugon, an AI assistant replies to Instagram direct messages and Facebook Messenger chats using your prices, services, opening hours and policies. It can book appointments into your calendar, take enquiry details, suggest products and flag the conversations a person should handle. Every thread lands in one inbox alongside your WhatsApp and website chats.
Before you get excited, the honest part: connecting Instagram to any automation tool runs through Meta's developer platform, and Klugon's Instagram setup includes Meta app review. This page explains what that involves, step by step, with links to Meta's own documentation, so you can decide whether it's worth it for your business.
What Instagram DM automation means for a small business
The phrase covers two quite different things, and it's worth separating them before you choose a tool.
DM replies vs comment-to-DM funnels
Comment-to-DM funnels are the marketing tactic you've probably seen: "Comment GUIDE and I'll send you the link." Software watches your comments for a keyword and sends a scripted DM to everyone who uses it. It's built for creators and brands running campaigns. Tools like ManyChat specialise in it; our ManyChat comparison covers where that kind of tool is the better fit.
Conversational DM replies are what a small business usually needs most: someone who answers the messages people actually send, in their own words, accurately. "Do you do lash lifts?" "Have you got the green one in a 12?" "Can I bring my dog in on Tuesday?" These aren't campaign responses. They're customer service and booking requests that happen to arrive on Instagram.
Klugon does the second. It replies to direct messages. It doesn't watch your comments, run keyword triggers, or send DMs to people who haven't messaged you. If comment funnels are your main goal, look at a dedicated tool. If you're drowning in real questions, keep reading.
Why DMs are hard to keep up with
Instagram DMs have a few habits that make them harder than email:
- They arrive in bursts. A post or a story that does well brings a wave of messages in a few hours, often outside working hours.
- They're short and vague. "Price?" under a story reply could mean any of six services.
- They arrive in two places. Messages from people who don't follow you can land in the requests folder, which is easy to miss.
- People expect a quick reply. On a chat app, a reply the next day feels slow, and the person has often moved on.
None of this is a reason to hire someone to watch Instagram all evening. It's a good reason to have something that answers the routine questions straight away and leaves the rest for you.
What the assistant does in Instagram DMs and Messenger
Klugon runs one AI assistant across every channel you switch on: Instagram, Messenger, WhatsApp, your website widget and, on the voice plan, the phone. You teach it once. If you've seen how Klugon's AI receptionist answers calls, this is the same assistant, writing instead of speaking.
Answering questions from your own information
The assistant answers from the material you give it: uploaded files (PDF, Word, plain text, Markdown or CSV), pasted text like your FAQ or aftercare notes, a single web page you import, or a scan of your homepage and up to seven key pages. You can also keep a library of links it's allowed to send: your booking page, price list, map, returns policy or a lookbook.
When a question isn't covered, it tells the customer someone will get back to them, and hands the conversation to your team, rather than inventing an answer. You'll find the gaps on the Insights page under "Questions without a good answer", which is the quickest way to improve it in the first month.
Booking appointments from a DM
For service businesses, this is the part that pays for itself. The assistant books into Klugon's built-in calendar, which knows:
- your opening hours and timezone
- each service and how long it takes
- which staff member does which service, so clashes are checked per person
- buffers between appointments, minimum notice and how far ahead people can book
- days you're closed
A customer who asks "anything Saturday?" gets real free times, picks one, gives a name and gets a booking reference in the chat. In DMs and Messenger the assistant can also move or cancel a booking when the customer asks. Your bookings list shows everything it booked, and you can edit or cancel from there.
If you live in Google, Outlook or Apple Calendar, or use Calendly, the assistant can also check those calendars and send Calendly links after you set them up, on the Business plan or above. It isn't automatic two-way sync. The AI appointment booking page goes through the calendar options properly.
Product questions and suggestions
Shops get a different kind of DM: "Is this still available?", "Do you have it in black?", "What's the price on the bag in your last post?". If you add your products to Klugon's catalogue, the assistant can search it, give prices and suggest related products. The catalogue is entered by hand and doesn't sync with an online store, so treat stock as "we listed it" rather than "it's on the shelf right now", and tell the assistant to hand over hold requests. The AI sales assistant page covers the selling side in more detail.
Taking enquiry details
Some DMs are the start of a bigger job: a bridal party, a corporate order, a custom piece, a photo shoot. The assistant can take the person's name, what they want, dates and contact details, and save them as a lead. Leads sit on the Leads page with a status (new, contacted, converted or lost), and you can export them. If you use HubSpot, Zoho, Salesforce or Pipedrive, the assistant can also send leads there once connected, on the right plan.
Facebook Messenger chatbot: the same assistant on your Page
Many small businesses get fewer Messenger messages than Instagram ones, but the people who use Messenger tend to be ready to book: they've found you on Facebook, often through a local group recommendation, and they want to know if you can fit them in.
Klugon's Messenger connection uses the same assistant, the same knowledge and the same calendar as Instagram. It replies to messages sent to your Facebook Page. Messenger has its own Meta setup, which is simpler than Instagram's in one respect: Meta's Messenger Platform overview says app review is "Not required if you only send and receive messages for your own Facebook Page."
So if Messenger is a meaningful channel for you, it's a sensible place to start, and you can add Instagram once its review is through.
One omnichannel inbox for small business
If you use more than one channel, the problem isn't only volume. It's that conversations are scattered: some in the Instagram app, some in Meta Business Suite, some in WhatsApp on someone's phone, some in a website chat tool. Nobody can see the whole picture.
Klugon puts them in one list. Instagram, Messenger, WhatsApp, the website widget and phone calls all appear in the same inbox, with:
- search across conversations
- a status filter: all, escalated, active or closed
- a channel filter, if you only want to see Instagram
- the full message history of each conversation
- internal notes your team can add
- transcript download and CSV export
On Instagram, Messenger and WhatsApp, your team can also reply to the customer from the inbox. The reply goes out from your Instagram account or Facebook Page, in the same thread the customer started.
What the inbox doesn't do yet: assign a conversation to a named team member, add labels, set priorities, or pause the AI while a person takes over. For a team of one to five people that's usually workable with a simple rule. For a large support team, it's a real gap.
What Klugon's Instagram DM automation won't do
So you can plan around them, here are the limits today:
- No comment replies or comment-to-DM triggers. It answers direct messages only.
- No outbound messages. It doesn't send broadcasts, promotions, "we miss you" messages or follow-ups. It replies when someone messages you.
- No reminders. There are no appointment reminders before a booking. The confirmation is the reply in the chat.
- No real-time transfer to a person. It flags the conversation and notifies your team through Slack or Zapier. A person then replies from the inbox.
- No email or push alerts. Notifications go to Slack or Zapier only.
- No pause on handover. The AI carries on replying if the customer keeps writing after it has flagged the conversation.
Setting up Instagram DM automation with Meta
Meta controls access to Instagram messages through its developer platform. Any tool that reads and replies to your DMs needs an app registered with Meta, permission to use the messaging API, and a token. Klugon doesn't yet have a one-click "Log in with Facebook" connection, so you create these in your own Meta developer account and paste them into Klugon.
That's real work. If you've never been inside the Meta App Dashboard, expect to spend an afternoon on it, and then to wait for Meta's review. We won't put a minute figure on it, because most of the time is Meta's.
What you need before you start
According to Meta's overview of the Instagram Messaging API:
- An Instagram professional account, either business or creator. If you're on a personal account, Instagram's guide to setting up a professional account shows how to switch. A private personal account becomes public when you switch.
- A Facebook Page linked to that Instagram account. Meta handles Instagram conversations for businesses through the Page.
- Admin access to the Page and to your Meta business portfolio.
The setup checklist
These steps follow Klugon's Instagram setup card and Meta's documentation. Meta renames menus often; if a label differs, follow Meta's current wording.
- Switch Instagram to a professional account and link it to your Facebook Page.
- Allow message access. In the Instagram app, go to Settings, then Messages and content (or similar), then Connected tools, and turn on "Allow access to messages". Without this, no tool can read your DMs.
- Create a Meta app at developers.facebook.com and add the Messenger product. Instagram messaging runs through it.
- Request the permissions Klugon's setup lists:
instagram_basic,instagram_manage_messages,pages_messagingandpages_manage_metadata. - Generate a Page access token for the Facebook Page linked to your Instagram account.
- Paste the details into Klugon. In Integrations, open Instagram DM, paste the Page access token and your app secret (found under App settings, Basic), and choose a verify token: any phrase of eight characters or more.
- Set the webhook in Meta. In your app's webhook settings for Instagram, enter the callback URL shown on Klugon's setup card, the same verify token, and subscribe to
messages. This is how Meta tells Klugon a DM has arrived. - Assign an assistant to the Instagram channel in Klugon.
- Test with someone who has a role on the app. Until Meta approves the app, only people with a role on it can message it and get a reply. Add a colleague as a tester in the app's Roles section, message your account from theirs, and check the reply, a booking and a handover.
- Submit for app review for
instagram_manage_messages. Once it's approved, your customers' messages get replies too.
Meta app review for instagram_manage_messages
This is the step that surprises people. In Klugon's setup, Meta app review for the instagram_manage_messages permission is required before the assistant can reply to the general public. Until then, the app is in what Meta calls Standard Access, and Meta's overview is plain about what that means: apps with Standard Access "can only interact with people who have a role on the app."
Meta's App Review for Instagram API page explains which access level an app needs, and what a submission includes. From that page, the main things to prepare:
- A screencast for each permission you request, showing how your app uses it from start to finish. For messaging, that means recording a DM arriving on your Instagram account and the reply going out.
- At least one successful API call with the permission, before you can request Advanced Access. Your own test message through Klugon does this.
- A description of how you use each permission. Meta says it won't approve a submission that requests permissions the app doesn't use, or uses in a way that doesn't match the allowed use. Only ask for what's in the checklist above.
- An English interface in the recording if possible, or captions explaining the buttons.
Meta's review times vary, and reviewers sometimes come back asking for a clearer recording. Plan for the possibility of a second submission. In the meantime, you can run Messenger and WhatsApp, and test Instagram with people on your app.
Messenger setup
Messenger is the shorter path:
- Create a Meta app (or use the same one) and add the Messenger product.
- In Messenger's webhook settings, enter Klugon's callback URL for Messenger and a verify token you choose, and subscribe to
messagesandmessaging_postbacks. - Generate a Page access token for your Facebook Page.
- In Klugon, open Facebook Messenger under Integrations and paste the Page access token, your app secret and the same verify token.
- Assign an assistant to the Messenger channel and send a test message to your Page.
Klugon checks the signature on every Messenger webhook against your app secret, so it only acts on messages that really came from Meta.
If an agency or freelancer runs your Instagram
Plenty of small businesses have a social media manager, a nephew or an agency who set up the accounts years ago. Before you start, find out who is an admin on the Facebook Page and who owns the business portfolio. The Meta app and the Page access token should belong to the business, not to a freelancer's personal Facebook account. If that person leaves and their access is removed, a token generated through their account can stop working, and the assistant goes quiet without anyone noticing.
A sensible arrangement: the owner, or a long-term manager, creates the Meta app inside the business's own portfolio and is an admin on it. The agency can be given a role on the app to help with the setup and the review submission. When they finish, check the roles list and remove anyone who no longer needs access. Keep a note of which account generated the Page access token, so you know where to look if replies stop.
Keeping the connection healthy
The most common reasons replies stop: the Page access token was revoked (for example, when a Page admin changed their Facebook password or left), message access was turned off in Instagram's settings, or the webhook subscription was removed when the app was edited. If replies stop, check those three first.
The messaging window on Instagram and Messenger
Meta limits when a business can message someone on Instagram and Messenger, much as it does on WhatsApp. Meta's Messenger Platform policy overview describes a standard messaging window that, "For most user actions, it lasts 24 hours" from the person's last message. Inside it, a business can reply freely. Outside it, Meta allows only certain tagged messages, under its own rules.
For Klugon, that means:
- The assistant replies straight away to each message, so it's always inside the window.
- When your team replies to a flagged conversation from Klugon's inbox, do it within 24 hours of the customer's last message. Klugon doesn't check the window before sending; if Meta refuses the reply, the inbox shows Meta's reason in plain words.
- Klugon doesn't use Meta's message tags, so there's no way to send a message from Klugon after the window closes. If a customer goes quiet for two days, you'd reply from the Instagram app or Meta Business Suite instead, under Meta's rules there.
Sample Instagram and Messenger conversations
These are sample conversations, written to show typical exchanges. The businesses and customers are made up.
Sample 1: a nail studio answering a story reply
A sample Instagram DM conversation. The studio, "Lacquer Room", is made up.
What happened: "price??" on its own is ambiguous, but it was a reply to a specific story, and the owner had described that set in the knowledge base with a name customers would recognise from the post. The removal policy came from the studio's FAQ. A tip: when you post a new set or offer, add a line about it to the knowledge base the same day.
Sample 2: a clothing boutique and a request to hold an item
A sample Instagram DM conversation. The shop, "Marlow & Pine", is made up.
What happened: the assistant found the product in the catalogue and gave the price, but the owner's instructions said "never promise a size is in stock; hand over hold requests". It captured the details as a lead and flagged the conversation. A staff member checked the rail, replied from Klugon's inbox 40 minutes later, and set the conversation to closed.
Sample 3: a dog groomer on Facebook Messenger
A sample Messenger conversation. The groomer, "Scruff & Shine", is made up.
What happened: the owner had told the assistant to hand over badly matted or anxious dogs rather than book them straight in, because those appointments need a groomer's judgement. The assistant gave the guide price, named the free slots and flagged the conversation. The groomer replied that afternoon from the inbox and booked the Wednesday slot.
What the AI hands to your team
On Instagram especially, some messages need a person: complaints made in public that move to DMs, influencer and collaboration requests, wholesale enquiries, anything sensitive about a treatment, and anyone who asks for a person. You list these in the assistant's instructions. The guide to handing calls and chats to staff has example wording.
When the assistant hands a conversation over:
- It tells the customer that someone from the team will reply.
- It marks the conversation "escalated", and the inbox shows a "Needs a person" banner. Filter by "escalated" to see only these.
- It notifies Slack or Zapier, if connected. The Slack integration posts the handover to a channel; Zapier can send it to the tool your team already uses.
- It adds it to the Insights to-do list under handovers waiting.
Your team member opens the thread, reads the full history and any details the assistant collected, and replies from Klugon's inbox. They can add an internal note ("refunded in store, Saturday"), and close the conversation when it's done.
Two things to agree with your team in advance. First, since the AI isn't paused, the person replying should keep an eye out for the assistant also answering if the customer writes again. Second, reply within 24 hours of the customer's last message, or Meta won't deliver it from Klugon.
Writing the assistant's instructions for DMs
DMs have their own register. A few things that help the assistant sound like your account:
- Match your posts. If your captions are relaxed and use first names, say so in the instructions. If you're a solicitor's practice with a Facebook Page, ask for a more formal tone.
- Name things the way your posts do. If customers know a set as "the chrome French", put that phrase in the knowledge base next to the service.
- Keep price answers specific. "From £48, depending on length" beats "prices vary".
- Update the knowledge base when you post an offer. A story about a mid-week discount will bring messages about it. If the assistant doesn't know, it will hand them all over.
- List what to hand over, in plain sentences: "Hand over any complaint." "Hand over collaboration and influencer requests." "Hand over any question about a reaction after a treatment."
- Add your links. Booking page, price list, location, and your returns policy for shops.
Then read the first week of conversations properly. You'll spot answers that are accurate but don't sound like you, and you can fix them in the instructions.
Pricing and a worked cost example
Instagram and Messenger are part of Klugon's AI social agent plan, which also covers WhatsApp. All prices are in US dollars; the full details are on the pricing page.
| Social plan | Monthly | Billed yearly (per month) | Conversations a month |
|---|---|---|---|
| Starter | $79 | $63 | 300 |
| Business | $249 | $199 | 1,500 |
| Scale | $449 | $359 | 3,500 |
A conversation is one customer on one channel, whatever the number of messages. The same person writing on Instagram and on WhatsApp counts as two. Instagram, Messenger and WhatsApp share the same allowance. If you reach 110% of it, new conversations stop; ongoing ones carry on.
Connecting Instagram or Messenger doesn't add a fee. Each WhatsApp number you connect adds $19 a month (Starter connects one, Business up to 3, Scale up to 10), covered on the WhatsApp AI chatbot page. Every plan includes unlimited assistants. Calendar and CRM integrations need the Business plan or above.
The trial is 7 days. A card is required at sign-up and is charged when the trial ends, unless you cancel first.
A worked example
Say you run a nail studio with three technicians. Assumptions:
- You get around 250 DM and Messenger conversations a month, more in the weeks before Christmas and in wedding season.
- Starter plan at $79 a month covers 300 conversations. In December you might go over, so either plan for the Business tier that month or accept that new conversations stop at 330.
- About 1 in 5 conversations is a booking request.
- Before the assistant, you reckon you lost a quarter of those because replies went out the next day.
That's 50 booking requests a month, of which about 12 used to go elsewhere. At an average of £40 a set, that's roughly £480 a month you weren't getting. Against $79 a month, the maths is clear, as long as your assumptions are roughly right. Check them against your own Instagram inbox for a month before you decide: count the booking requests and how quickly you answered them.
Separately, if you're on Business or above for calendar sync or CRM, the plan is $249 a month and covers 1,500 conversations, which is plenty for most single-site businesses.
How to tell whether it's working
After a month, look at:
- Conversations by channel. The Analytics page shows how your conversations split across Instagram, Messenger, WhatsApp and your website, and the busiest hours. If most Instagram volume lands in the evening, that's where the assistant is doing work nobody did before.
- Bookings. Count the bookings made in DMs. The Analytics page shows bookings (not counting cancellations) for any date range, compared with the previous period.
- Handled without handover. Analytics shows the share of conversations the assistant handled without handing over. A low figure early on usually means gaps in the knowledge base.
- Questions without a good answer. Insights lists them. Fix the top few every week.
- Most asked questions. The last 30 days, on the Insights page. Use them to write better captions and highlights, too.
And read a sample of conversations each week. A number tells you how many; a transcript tells you whether customers got a good answer.
Who it suits, and who should look elsewhere
It suits service businesses that take bookings and get a steady stream of DMs: nail, lash and brow studios, hairdressers, barbers, beauty and aesthetics clinics (for admin questions, not treatment advice), personal trainers, groomers, tattoo studios, photographers. The beauty studios page goes into that use case. It also suits small shops whose DMs are mostly price, size and opening-hours questions; see the clothing boutiques page.
It's less suited to creators and brands whose main need is comment-to-DM campaigns and broadcasts, or to large support teams that need assignment, tags and a pause button. It's also not the right tool for anyone who can't get through Meta's setup: if your Instagram account isn't a professional account linked to a Page you control, sort that out first.
If your website gets as many questions as Instagram, the same assistant runs as an AI chatbot for your website, and those chats land in the same inbox.
Frequently asked questions
Is Instagram DM automation allowed by Meta?
Yes, through Meta's official Instagram Messaging API, which requires an Instagram professional account linked to a Facebook Page and an app with the messaging permissions. Klugon uses that API, through a Meta app you create and control.
Why does Klugon need Meta app review for Instagram?
Klugon's Instagram setup needs the instagram_manage_messages permission at a level that lets the assistant reply to the public. Until Meta approves the app, only people with a role on it can message your account and get an AI reply, so the assistant can't serve customers yet. Meta's App Review for Instagram API page explains the process.
How long does Meta's app review take?
Meta doesn't promise a fixed time, and it varies. Some submissions are approved on the first try; others come back asking for a clearer screencast. While you wait, you can connect Messenger and WhatsApp and test Instagram with people on your app.
Does Messenger need app review too?
According to Meta's Messenger Platform overview, app review is not required if your app only sends and receives messages for your own Facebook Page. You still need a Meta app, a Page access token, your app secret and a webhook pointed at Klugon.
Can the assistant reply to comments on my posts?
No. Klugon answers direct messages on Instagram and messages to your Facebook Page on Messenger. It doesn't reply to comments or send DMs triggered by comment keywords.
Can it book appointments from an Instagram DM?
Yes. It books into Klugon's built-in calendar using your opening hours, service lengths and staff, and gives the customer a booking reference in the chat. It can also move or cancel a booking if the customer asks in the same channel.
What happens when a customer asks for a person?
The assistant tells them someone will reply, marks the conversation as needing a person, and notifies your team in Slack or Zapier if you've connected them. Your team replies from Klugon's inbox within 24 hours of the customer's last message. The AI isn't paused in the meantime.
Do I see Instagram, Messenger and WhatsApp in one place?
Yes. Klugon's inbox lists conversations from Instagram, Messenger, WhatsApp, the website widget and phone calls together, with search, status and channel filters. You can reply to Instagram, Messenger and WhatsApp conversations from it.
Does it cost extra to add Instagram or Messenger?
No. Both are included in the AI social agent plan, from $79 a month for 300 conversations, shared with WhatsApp. Only a WhatsApp number adds a fee, of $19 a month.
Getting started
If your DMs are mostly the same twenty questions and a stream of booking requests, an assistant that answers from your own information and books into your calendar will save you real evenings. Start with Messenger, which is quicker to connect, submit Instagram for Meta's review, and use the waiting time to write a good knowledge base. You can sign up for the 7-day trial and try the assistant in your browser first.