Answers
ChatGPT is safe for business data when accessed through a read-only connector that limits what the model can see and does not allow training on customer data. CorpusIQ exposes Gmail, Shopify, QuickBooks, and 22 other tools to ChatGPT through read-only OAuth. Customer data is not used for model training. No files are stored on CorpusIQ servers. Every answer is cited so the operator can verify the source. CorpusIQ maintains a SOC 2 aligned posture and is CASA Tier 2 certified by DEKRA.
Related pillar: Security and compliance posture
Three controls: read-only access by default (the assistant cannot write or delete), no model training on customer data, and cited answers so every claim is auditable. CorpusIQ delivers all three by design. ChatGPT Enterprise and Team add admin controls on top.
Each of the 25+ connectors uses an OAuth scope that grants read access only. The assistant can list, read, and quote from records you have authorized. It cannot send, modify, move, or delete anything. Revoke any connector in one click; that ends ChatGPT's access to that tool immediately.
CorpusIQ's contractual position is that customer data is not used to train any model. OpenAI's commercial terms for ChatGPT Enterprise and Team likewise state that business inputs and outputs are not used to train OpenAI models. The CorpusIQ + ChatGPT pairing is, in practice, doubly opted out.
Records are read on demand to answer a question, then released. CorpusIQ does not retain customer files on its servers. There is no index, no embedding store, no cached copy. Each question reads fresh from the connected tool.
Connect only the tools you want exposed. Within each tool, the OAuth scope is fixed to read-only; you cannot reduce it further from the OAuth side, but you can scope a question to a folder, project, or client by naming it in the prompt. To remove a tool from the assistant's reach, disconnect it in CorpusIQ.
25+ read-only connectors. ChatGPT, Claude, and Perplexity. Solo $29.95 a month. 30-day free trial.
