I gave ChatGPT my Instagram. It built a content analyst.

4.3 million views converted about fourteen times worse than a 122,000-view AI experiment. Getting to that sentence took 60 reels, 60 sets of Insights, 59 transcripts and one patient spreadsheet. ChatGPT Work assembled all of it.

reels

6.6M

views

0.45

median follows per 1K

I did not install five tools and wire them together. I told ChatGPT Work the outcome, pointed it at my signed-in browser and a Google Sheet, and let it find or install the missing pieces in isolated local environments. When something needed a login, it opened a local prompt and I typed the credentials there, never in chat.

The Stevie Builds Instagram reel grid with view counts — Starting point. A grid, some view counts, and no way to compare videos.

One request, six jobs

Browse

Open the reel grid, find the new shortcodes, collect the visible metadata.

Inspect

Open each reel's Insights and record the numbers views alone hide.

Download

Save each MP4 locally, one at a time.

Transcribe

Run Parakeet locally, one video at a time. Nothing leaves the machine.

Catalogue

Write covers, transcripts, hooks, types and conversion into Sheets.

Explain

Separate what gets distributed from what earns the follow.

Browse the account and capture the Insights

The browser pass deduplicated reels by shortcode, saved each cover and opened View Insights. It recorded views, follows, likes, comments, shares, saves, reach, engaged accounts and profile activity. Those are the numbers that separate entertainment from follower growth.

An Instagram reel page showing the View Insights button — The escape hatch. Open Insights by hand if Instagram refuses to route automatically.

Tell it the outcome

This is the reusable version of the request. Change the account and the sheet, then let the agent work out the how. The constraints matter: use the signed-in browser, keep credentials out of chat, download sequentially, transcribe one video at a time.

reel-catalogue-prompt.txt

I want you to build and maintain an Instagram Reel content tracker for my account.

Use my existing signed-in Instagram session and Google Sheet. Work incrementally: compare Reel URLs or shortcodes first, then add only reels that are not already in the tracker.

For every new reel:
1. Capture the reel URL, shortcode, posted date, caption, current view count and a clear working title.
2. Save the cover as a local PNG or JPEG and embed the actual image in the sheet — do not use an externally hosted IMAGE formula.
3. Open View Insights and record views, follows, likes, comments, shares, saves, accounts reached, accounts engaged, interactions and profile activity.
4. Calculate followers per 1,000 views.
5. Download the video locally. Install an isolated copy of yt-dlp if needed. Do not extract my browser cookies.
6. Transcribe locally with Parakeet MLX. Process exactly one video at a time so the computer does not run out of memory.
7. Store the transcript, transcript status, a consistent content type and the actual opening hook in the sheet.
8. Verify the new rows and report any reels whose Insights or audio could not be retrieved.

Once the catalogue is complete, analyse the spreadsheet only. Compare reach, follower conversion, shares per 1,000, saves per 1,000, content type and hook. Separate content that earns broad distribution from content that gives viewers a reason to follow.

You probably do not need to install anything

ChatGPT Work can create isolated Python environments, install the downloader and the transcription package, check the media and write the helper scripts. You step in for a login, a CAPTCHA, or a decision that changes what it can reach.

If you want to see the stack, the setup below reproduces the local pieces. Parakeet MLX runs well on Apple silicon. FFmpeg inspects and extracts audio. yt-dlp downloads public reel media. Instaloader helped with an authenticated profile pass, but its direct-reel endpoint was unreliable during this run.

optional-local-setup.sh

# Optional: let ChatGPT Work run these for you
python3 -m venv ~/.local/share/reel-lab
~/.local/share/reel-lab/bin/python -m pip install --upgrade pip yt-dlp instaloader

# Local Apple-silicon transcription
python3 -m venv ~/.local/share/parakeet-mlx
~/.local/share/parakeet-mlx/bin/python -m pip install --upgrade pip parakeet-mlx

# macOS media inspection / audio extraction
brew install ffmpeg

Download, transcribe, structure

Every video went into a dated local folder. Parakeet produced 55 usable transcripts. Four reels had no detectable speech and one had no audio stream. Those gaps were recorded, not hidden.

The completed Reel Catalogue Google Sheet with covers, transcripts and metrics — The finished dataset. Embedded covers, transcripts, hooks, content types and follower conversion.

Viral and follow-worthy are different products

Reach engine · workplace satire

57.3K median views

Recognisable workplace pain is easy to send to a colleague. The joke is complete once watched, so there is little reason to follow. 0.06 follows per 1K.

Conversion engine · AI-agent demos

4.87 median follows per 1K

A specific experiment gives novelty, proof and an unfinished story. People follow because they expect another test tomorrow. 11.5K median views.

A chart comparing reel views with followers per thousand views — Top right is the goal: enough reach to find new people, enough specificity to earn the follow. One reel sat there.

Saves were the strongest signal

Saves / 1K

0.64

Interactions / 1K

0.64

Likes / 1K

0.50

Comments / 1K

0.41

Shares / 1K

0.36

Spearman correlation with followers per 1,000 views. Not causation, but a useful pattern: content people want to keep is close to content people follow for.

The reel blueprint

Start with a familiar conflict. A claim or a workplace pain that lands before the viewer knows who you are.
Make it your experiment. Something you built, tested or made the agents do. Not AI news.
Show the proof. The tool, the result, the spreadsheet or the failure, on screen.
Leave one loop open. Say what you will test next so following has a payoff.

The reel I would make about this

The hook carries the scale, the automation and the counter-intuitive result in one breath. The middle shows the process. The ending promises the next twelve experiments instead of a generic follow ask.

chatgpt-reel-analysis-script.txt

[0–4s — hook]
I gave ChatGPT Work my Instagram account. It watched 60 of my reels, downloaded every video, transcribed them locally — and then told me why 4.3 million views barely converted.

[4–11s — show the profile grid]
I gave it one spreadsheet and my signed-in browser. It opened every reel, grabbed the cover, caption, view count and the full Insights.

[11–19s — show View Insights]
That meant follows, shares, saves, reach — everything Instagram normally makes you inspect one video at a time.

[19–27s — show downloads / transcription]
Then it installed its own downloader, saved all 60 videos, and ran a local transcription model over them. One video at a time, because I nearly cooked my computer the first time.

[27–36s — reveal the sheet]
It turned the whole account into this: searchable transcripts, hooks, content types, covers and follower conversion for every reel.

[36–48s — the result]
And this was the surprise. My 4.3-million-view workplace joke earned 0.36 followers per thousand views. This 122-thousand-view AI-agent experiment earned 4.87. Nearly fourteen times better.

[48–57s — takeaway + CTA]
Funny, relatable videos get shared. Specific experiments make people want to see what you build next. I put the full workflow, prompts and tools in the guide in my bio. Follow — because I'm testing the blueprint on the next twelve reels.

The next experiment

Five concrete AI experiments, three workplace sketches, two AI tutorials, one personal story and one internet find. Measure reach, saves and follows per 1,000 separately. Then run this workflow again.