Once you’ve connected Instagram to ChatGPT (see: How to Connect Instagram to ChatGPT), the real value comes from asking the right questions. Below is a ready-to-use prompt library covering the five areas that actually move the needle: content format, timing, audience health, Story performance, and creative strategy.
These aren’t generic ‘write me an Instagram caption’ prompts. Each one queries your actual connected data, real fields like media_saved, media_reach, and story_exits, so the answers reflect what’s actually happening in your account, not a generic best-practice guess.
Content format analysis: Reels vs. Carousels vs. single images
Gut feeling is often wrong about which format performs. Different formats drive different types of engagement, and not all metrics matter equally.
Format comparison by the signals that actually matter
Group all posts from the last 90 days by media_type (IMAGE, VIDEO, CAROUSEL_ALBUM). For each format, calculate averages: - media_reach (raw audience size) - media_reach / followers_count × 100 (reach rate) - media_saved / media_reach × 100 (save rate) - media_shares / media_reach × 100 (share rate) - media_engagement / media_reach × 100 (true engagement rate) 1. Which format has the highest reach rate? 2. Which format drives the most saves per impression? 3. Is there a format that performs worse on raw reach but better on saves and shares?
Reel performance: what makes a video keep getting played
For all Reels posted in the last 60 days, show: - media_impressions, media_reach, media_shares - Replays vs. initial plays ratio 1. Which Reels have the highest replay rate? 2. Do my highest-replay Reels share common characteristics in captions or posting times? 3. Compare share rate of Reels vs. Carousels — which gets shared more?
Timing and scheduling: when your audience is actually watching
Generic “best times to post” advice is based on averages, not your specific audience.
Finding your real best posting time
Using my Instagram data for the last 30 days, show audience activity by: - Hour of day - Day of week 1. Which 3 two-hour windows have the highest follower activity? 2. Compare this against the timestamps of my last 30 posts. 3. Am I posting during peak windows, or before/after my audience is online? Suggest the best posting windows based on this.
Posting frequency vs. reach — is more always better?
Group my posts from the last 6 months by week. For each week show: - Number of posts, average media_reach, average media_engagement, followers_count 1. Is there a correlation between posting frequency and average reach? 2. In weeks I posted more than 5 times, did per-post reach go up or down? 3. Is there a frequency sweet spot where both reach and engagement beat the period average?
Audience health: are you attracting the right followers?
Demographic drift
Using audience_gender_age, audience_city, and audience_country, compare the last 30 days vs. the prior 30 days: gender breakdown, age brackets, top 5 cities, top 5 countries. 1. Has my audience composition shifted materially? 2. Does any bracket growth align with recent content changes? 3. Are there new cities/countries gaining share worth targeting?
Profile visit to follow conversion
For the last 30 days, show daily profile_views and net follower change. Calculate: net new followers / profile_views × 100. 1. On high-profile_views days, what was posted that day or the day before? 2. On high-views/low-conversion days, what content pattern shows up?
Story performance: where viewers leave and where they lean in
For multi-frame stories in the last 30 days, show per frame: story_impressions, story_reach, story_exits, story_taps_forward, story_taps_back, story_replies. 1. For stories with exits above 30%, at which frame does the spike occur? 2. Which frames generated the most taps_back, and what was in them?
Creative strategy: what to post and how to make it work
Identify top-performing content
Analyze my posts from the last 30 days. Identify the top 10 by saves and shares. What topics, formats, and content types do they share?
Generate new content ideas
Generate 10 new Instagram post ideas based on my top-performing content themes. Include format (Reel, Carousel, Image) and a short concept for each.
Write high-performing captions
Write 3 Instagram captions based on my top-performing posts. Use similar tone, structure, and hooks. Optimize for saves and shares.
Build a content plan
Create a 2-week Instagram content plan based on my top-performing content. Include post types, topics, suggested hooks, and text/image ideas. Plan both posts and Stories.
Related reading
Once you’ve found your best-performing content here, two natural next steps:
- Turn a winning post into a live campaign with ‘How to manage Instagram ad campaigns from ChatGPT‘
- Check how your numbers stack up against your competitors with ‘How to benchmark your Instagram against competitors using ChatGPT‘.
🚀 Running a Facebook Page too? See ‘How to Connect Facebook Page Insights to ChatGPT‘ for the same style of organic content analysis on Facebook.
FAQs
How is this different from a generic "ChatGPT prompts for Instagram" list?
Generic prompt lists ask ChatGPT to invent captions or ideas with no data behind them. These prompts pull your actual post, Story, and audience metrics first, so recommendations are grounded in what’s actually performing, not a general best-practice guess.
How far back can ChatGPT analyze my Instagram data?
As far back as your connected Instagram Insights history goes. Most prompts above default to 30, 60, or 90-day windows, but you can ask for any range.
Can ChatGPT write captions in my own brand voice?
Yes, ask it to base captions on your top-performing posts and it will match their tone, structure, and hooks rather than generating generic copy.
Do these prompts cover Instagram Stories, or just feed posts?
Both. The Story performance prompts use fields like story_exits, story_taps_back, and story_replies to analyze frame-by-frame retention, separately from the feed-post and Reel prompts.
Which Instagram metrics can I get through Windsor that are hard to get in native Insights?
Windsor surfaces the same underlying data as your Professional Dashboard, but makes cross-post and cross-time-period analysis possible: frame-by-frame Story retention (story_exits, story_taps_back), reach rate relative to your follower count, and save or share rate as a percentage of reach, things Instagram’s own interface shows one post at a time rather than as a queryable dataset.
Windsor vs Coupler.io

