Capturing one ChatGPT conversation at a time is fine for a spot check. For a category study you need a hundred. Bulk Prompt Analysis drives ChatGPT through a prompt list on its own and captures every answer into your history, ready for the Bulk Analysis dashboard.
Purpose
- Run a whole prompt list through ChatGPT unattended and capture every answer.
How to Open
- Popup nav →
AI Insights→ Bulk analysis → the Run segment. - You need to be signed in to ChatGPT. The run drives a real ChatGPT session in a tab, using your own account.
Starting a Run
- Paste your prompts into the box, one per line. The counter shows how many were detected.
- Optionally assign the run to a project, so every capture lands grouped by client or campaign.
- Click Start.
Input format. One prompt per line is the rule. A line made entirely of quoted CSV cells ("first prompt","second prompt") is split into one prompt per cell, so a row copied out of a spreadsheet works. A plain line containing commas is treated as a single prompt, not split.
Duplicates are kept on purpose. ChatGPT does not answer the same prompt the same way twice, so repeat-sampling a prompt is a legitimate method. The counter tells you how many duplicates it found and runs them as-is rather than quietly removing them.
Saved lists. Name a prompt list and save it, then load, rename or delete it later. Useful when you run the same category set every month.
While It Runs
Prompts are processed one at a time, in order, in a single ChatGPT tab. There is no parallelism, deliberately: concurrent tabs would trip rate limits almost immediately.
The Progress panel shows each prompt with a status: Pending, Done, Error, or Timed out.
Controls:
- Pause and Resume.
- Stop ends the run.
- Clear empties the progress list.
- Retry unfinished re-queues everything that did not complete.
A run survives closing the page. State is saved as it goes, so you can shut the tab and pick the run back up.
If you open Bulk analysis in a second tab while a run is going, the second tab will tell you the batch is controlled by another tab. Only one tab drives a run at a time.
Limits and Safety Rails
These exist because ChatGPT will rate-limit an aggressive run, and a half-captured batch is worse than a slow one.
- Over 50 prompts triggers a confirmation warning that large batches take a long time and may hit rate limits.
- 180 seconds is the hard timeout per prompt.
- After an answer finishes streaming, the extension waits for ChatGPT to persist the conversation, polling for about 16 seconds before giving up on that item.
- Auto-pause kicks in after 3 consecutive hard failures, or 5 consecutive answers that never persisted. Both usually mean a rate limit or a signed-out session. The run pauses rather than burning through the rest of your list producing nothing.
- Failed items get one automatic retry, and only when the failure happened before the prompt was submitted. An answer that was submitted is never silently sent twice.
- Pacing is deliberately conservative: a short settle after opening a new chat, and a gap between items.
Each captured answer is also checked against the prompt that was submitted before it is saved, so a mis-timed capture does not get attributed to the wrong prompt.
Outputs
Every captured prompt becomes a normal entry in your AI Insights history, with the full capture depth: fan-out queries, sources, sourcing pipelines, products, carousels, entities and reasoning. From there:
- Open any single conversation in the AI Insights tab.
- Read them in aggregate in the Bulk Analysis dashboard.
- Export any section as CSV, or the whole set as JSON.
Workflow Wins
- Turn a keyword list into an AI-visibility study without sitting there pasting prompts.
- Re-run the same saved list monthly and watch the domain mix move.
- Sample one high-value prompt twenty times to see how stable the answer actually is.
Recommended Plays
- Category sweep: 50 buying-intent prompts for one category, assigned to a project, run overnight.
- Variance check: the same prompt 20 times, to find out whether a citation is reliable or a coincidence.
- Before and after: run a list, publish your changes, run the identical saved list a month later, compare.
Tips & Troubleshooting
- Run stops early? Check whether it auto-paused. Three failures in a row usually means ChatGPT rate-limited you or your session expired. Sign back in and Resume.
- Items marked Timed out? The answer took longer than 180 seconds, or never got persisted in time. Use Retry unfinished.
- Nothing captured but the answers appeared? Confirm you are signed in and that AI Insights captures a single conversation normally first.
- Keep the batch modest to begin with. Twenty prompts will tell you whether your list is well-formed before you commit to two hundred.
- Do not close the browser mid-run, though you can close the Bulk analysis tab itself. The run needs the browser open to drive ChatGPT.
Frequently Asked Questions
AI Insights: Common Questions
The Brands card shows no mentions. What’s wrong?
Nothing is scored against your brand list until you press Reprocess all conversations in the Bulk Analysis dashboard.
Add your brands and competitors first, then Reprocess. It re-scans everything you have already captured against the current list, so you do not need to re-run any prompts. If counts still look low, add more aliases, the misspellings, the legal entity name, the bare domain, and Reprocess again.
Why did my bulk run stop part-way through?
It almost certainly auto-paused, which is deliberate. A run pauses after 3 consecutive hard failures, or after 5 consecutive answers that never got persisted. Both usually mean ChatGPT rate-limited you or your session expired. Sign back in and press Resume rather than restarting the batch.
Items marked Timed out exceeded the 180-second per-prompt limit. Use Retry unfinished to re-queue just those.
Keep first batches small. Twenty prompts will tell you whether your list is well-formed before you commit to two hundred.
What is a sourcing pipeline, and why does it matter?
It is the retrieval path ChatGPT used to fetch a given source. ChatGPT records one of four and the extension reads the label directly: Bright Data and Oxylabs are commercial web scrapers, Labrador is an allowlist of licensed publishers such as Reuters, WSJ, Wikipedia and arXiv, and Open web / SERP is the search baseline.
It matters because it tells you which lever would actually move you. If answers in your category come mostly through the scrapers, being fetchable and answering the question in the page body is the work. If they come mostly through Labrador, you will not optimise your way in and the lever is coverage by those publishers instead.
Can I capture an old ChatGPT conversation, or only a live one?
Old conversations work. Extraction reads the conversation from ChatGPT’s own backend using the session you are already signed in with, so you can open a chat from last month and capture it in full.
Two things are the exception. Turn intent and live merchant offers are only caught while the answer is streaming, so they will not appear on a conversation you reopen later.
Gemini is different: its capture listens to live network traffic. Reopening an old Gemini conversation produces no traffic to listen to, so it falls back to reading the page and finds fewer queries. Capture Gemini conversations while you are having them.
Which Google domains does AI Overview capture work on?
Currently 17: google.com, .co.uk, .nl, .de, .fr, .es, .it, .ca, .com.au, .co.jp, .com.br, .co.in, .be, .pl, .se, .at and .ch.
On any other Google domain the AI Insights tab will not pick up an Overview, which is the most common reason it looks empty on a SERP that clearly has one. This is being widened to every Google country domain in the next release.