How to set up Instagram AI DM replies with your own FAQ
Short answer
In Instagram AI Systems, give DM Auto Reply your approved FAQ, product details, prices and delivery terms, then define a lead goal and qualification rules. Set reply length, history depth, threads and replies per run, and review how message requests go to General, Primary or spam deletion before enabling the activity.
Video demonstration
6 min 3 secOpen the full demo page
Key points
- DM Auto Reply [AI Powered] answers incoming DMs and message requests from your knowledge base and lead goal.
- A knowledge base supplies the facts; a lead goal supplies the next step. Keep these responsibilities separate.
- Worthwhile leads are accepted into General, requests needing a human into Primary, and spam is deleted.
- The website demonstration uses sample data. Configure and review the settings in your own installed app.
Locate DM Auto Reply in the control panel
The AngeryWeb Insta control panel separates Settings, Schedule and Activity Settings. Settings holds account and API credentials. Activity Settings is where you enable DM Auto Reply and tune its activity controls; Schedule handles the pacing shared with the other modules.
Use the control-panel demonstration below to locate these controls, then configure them in your installed Windows app.
Prepare a knowledge base with approved facts
Collect the answers a person on your team would be allowed to give: what you sell, current prices and currency, availability, delivery details, terms and the correct purchase or contact link. Replace every bracketed placeholder in this starter template before using it.
This is a writing template for your knowledge base, not a settings import file. Fill it with your current product facts and update it when your offer changes.
Plain-text template. Copying does not submit the template text.
BUSINESS Name: [your business name] What we offer: [approved product or service description] PRODUCT FACTS Price and currency: [current approved price] Availability: [confirmed availability or how a person can check] Delivery: [regions, method and confirmed timing] Terms: [approved terms and relevant link] Purchase or contact link: [official link] FAQ Q: [a real customer question] A: [your approved answer] Q: [another real customer question] A: [your approved answer] REPLY GUIDANCE Tone: [your normal customer-service voice] Use only the facts above. Do not invent stock, discounts or delivery dates. If information is missing, say it needs confirmation by a person. Do not claim an order, payment or refund has been processed. HUMAN HANDOFF Topics needing a person: [refunds, legal questions, complaints and your other topics] Approved handoff wording: [a short message your team can honor]
Set a lead goal and qualification rules
The lead goal tells the AI what to steer a useful conversation toward. For example: answer the visitor's product question using the knowledge base, then offer your purchase link if they want to buy.
Use the qualification filters for genuine business intent and likely spam. Refund, legal and complaint keywords are part of the described human-handoff workflow.
| Message request | Described routing | What to review |
|---|---|---|
| A lead worth replying to | Accept into General | The reply uses approved facts and a relevant next step. |
| A request needing a human | Accept into Primary | Your team can find it and follow up. |
| Likely spam | Delete | Qualification filters do not misclassify legitimate questions. |
Tune reply length, history and per-run limits
- Reply length limits the size of an answer. Keep enough room for a useful answer without repeating your entire FAQ.
- Conversation history depth controls how much context is considered. Review follow-up questions as well as first messages.
- Threads per run and replies per run control the workload of a pass. A checked thread is not necessarily a sent reply.
- The unread-or-changed option narrows which threads are opened. An unchanged conversation may not need another response.
- Randomized minimum waits, follow-up timing and occasional deferred replies affect when an answer is sent. The shared scheduler also has sleep hours and per-activity intervals.
Review a small set of representative conversations
Use the module's test run and review the reply and request routing against the examples in this table. These examples help you check your configuration using the facts you supplied.
For a reply-content change, edit the knowledge base or instructions. For a routing change, review the qualification and handoff rules.
| Test message | Expected review outcome |
|---|---|
| How much does it cost? | The reply matches the exact price and currency you supplied. |
| Do you deliver to my location? | Only confirmed regions and terms are used; missing information is not invented. |
| Can I have a discount you did not list? | No unapproved discount is promised. |
| I want a refund / I have a complaint | The configured human-handoff path is used. |
| A follow-up to an earlier question | The answer uses the conversation context without repeating an unrelated pitch. |
Keep replies and human follow-up in sync
Check General for active leads and Primary for requests that need a person. Use Overview and Recent activity to review what ran, including deferred replies.
Refresh the knowledge base whenever prices, delivery terms, availability or links change. Recheck your sample conversations after an edit. For timing across all five modules, continue to the Instagram automation scheduler guide.
Explore the Instagram AI demonstration
The existing AngeryWeb Insta control panel and feature demonstrations use sample data. They illustrate the app’s described workflows; they are not screenshots, recommended account limits or activity on your account.
App documentation
Product descriptions checked on .
The feature names, descriptions and controls in this guide follow the Instagram AI Systems product page and its AngeryWeb Insta control-panel demonstration.