Your audience already tells you what to write about. Collect the questions people repeat in comments and DMs, group them by topic with AI, and turn the top three into posts that end with a clear action. Questions asked right before a purchase bring the best leads.
TLDR: Pull one month of comments and DMs, let AI group them by topic and intent, write one answer post per high-intent topic, and add a keyword call to action. Track which posts bring replies and leads, then make more of them.
Gather the raw material first
Open the last 30 days of comments and DMs across your accounts and copy every question into one list. You do not need a special tool. A shared inbox helps because it puts everything in one place.
Mydrop's social inbox pulls comments and messages from your connected accounts into a single view, so the list builds itself as you reply. For a longer view of what your audience says outside your own posts, monitoring catches mentions of your brand and keywords.
Let AI group, you decide
Paste the questions into an AI tool and ask it to group them by topic and mark which look closest to buying. This is where AI earns its place, because sorting 300 comments by hand is slow and dull.
Then take the output with caution. AI groups well, but it does not know that "how do I cancel?" is a churn signal and "how much for 3 locations?" is a sales signal unless you tell it. Review the groups yourself and pick topics based on your business.
Rank by intent, not by volume
Not every question is a lead. Sort them like this.
| Question type | Example | Intent | Post to write |
|---|---|---|---|
| Price or plan | "How much for a team of five?" | High | A plain explanation of how pricing works |
| Comparison | "Is this better than X?" | High | An honest comparison, with limits |
| Setup | "How long does it take to start?" | Medium | A step-by-step walkthrough |
| General curiosity | "What does this do?" | Low | A simple explainer |
| Complaint | "My post failed to send" | Support | Reply in the inbox, not a post |
Write posts first for the top two rows. Our guide to turning low-engagement comments into community leads has a three-tier triage for the rest.
Write the post as an answer, then add one action
A strong lead post has three parts:
- The question, in the audience's own words. It tells the reader this post is for them.
- A direct answer. Give the real answer in the post, not a teaser.
- One next step. Ask readers to comment a keyword to get a guide or link.
Trigger-word automation can reply to anyone who comments your keyword with a message, so the action does not depend on someone being online. Keep the keyword simple and tell people what they will receive.
Use AI captions to draft the post from the question list, then rewrite the first line yourself. A hook in your own voice is what keeps a post from sounding like every other brand.
Track which topics produce leads
After each post, record three things: comments, keyword replies and any actual conversations that began in DMs. A post with modest likes but ten keyword replies is a winner for leads.
In Mydrop's data, a typical Instagram post earned a 3.1% engagement rate, so do not judge a lead post by likes alone. Judge it by the people who asked for more. If you want to find prospects beyond your own audience, see using social listening to find new customers.
Keep a monthly loop going
Make this a monthly routine, not a one-off project:
- Week 1: gather and group the month's questions.
- Week 2: write and schedule the top three answer posts.
- Week 3: reply to every keyword comment and DM promptly.
- Week 4: count leads per topic and choose next month's topics.
Replies matter as much as posts. A lead that waits two days for an answer is often gone. Slow review also costs leads, which is why a slow approval process hurts conversions.
Next step: copy the last 30 days of questions into one list today and mark the three closest to buying. Then start free with Mydrop to bring comments and DMs into one inbox and schedule the answer posts.
Google review
Trustpilot review
Product Hunt comment
