AI governance for ChatGPT
ChatGPT at work: redaction and visibility
PromptWall redacts sensitive data in every ChatGPT prompt before it reaches OpenAI. NetLens shows how the team is really using it. The AI Control Panel owns every seat.
Real usage
What people paste into ChatGPT
The patterns we see most often once a team adopts ChatGPT for real work.
Customer records for a summary or reply
An employee pastes a customer email, support transcript, or CRM export and asks ChatGPT to summarise it or draft a reply.
API keys and .env snippets while debugging
Debugging help requests routinely include real credentials copied straight out of a terminal or config file.
Internal documents for a rewrite
An internal RFC, a chunk of a private repo, or a draft contract, pasted in and asked to be polished or explained.
Signed-out and temporary chats
ChatGPT supports temporary and unauthenticated chats that don't tie to a corporate identity. PromptWall covers both, not just signed-in sessions.
What your employee typed
Using the attached payroll_Sept.xlsx, check why Asha Nair's salary of ₹92,500 did not reach account 50100234567891 (IFSC HDFC0001234).
What ChatGPT received
Using the attached payroll_Sept.xlsx, check why [PERSON_1]'s salary of [AMOUNT_1] did not reach account [BANK_ACCOUNT_1] (IFSC [IFSC_1]).
Coverage
How NexusNest covers ChatGPT
PromptWall
- Names, emails, phone numbers and other personal data in pasted customer records
- API keys, tokens, passwords and connection strings
- Source code, internal project names and other confidential business content
NetLens
NetLens shows how many people use ChatGPT, what they're using it for by topic, and how often PromptWall redacted something - built from redacted text only, never from anyone's raw prompts.
Learn moreAI Control Panel
The AI Control Panel buys and assigns ChatGPT seats centrally. Employees sign in without ever seeing a password, and offboarding is one revoke.
Learn moreFAQ
ChatGPT, common questions
Does it work with the ChatGPT desktop app?
Yes. The agent covers the ChatGPT macOS and Windows desktop apps at the device level, not just the browser, and the same coverage applies to ChatGPT Voice.
What about ChatGPT temporary chats and the signed-out flow?
Both are covered - signed-in conversations and the temporary or unauthenticated flow. PromptWall redacts sensitive spans either way, before the request leaves for OpenAI.
Does this affect Custom GPTs or file uploads?
File uploads are covered. Spreadsheets, PDFs and images attached to a conversation are redacted the same way as typed prompts before they reach ChatGPT. Custom GPTs receive the same redacted text.
Will OpenAI treat a redacted prompt as broken?
No. The redacted prompt is plain text with placeholders in place of the sensitive spans. ChatGPT answers from the surrounding structure, so the response still works.
Does the AI Control Panel replace ChatGPT Team or Enterprise?
No - it sits alongside whichever ChatGPT plan the company is on and centralises who has a seat, what it costs, and how it's revoked. Redaction and NetLens work the same regardless of plan.
Start with one team. Prove control before you scale AI.
Start with one team. See exactly how your company uses AI.