OpenAI won't let some customers export their chats, but this tool will

The article highlights that OpenAI’s ChatGPT Business/Enterprise customers cannot export workspace chats via the standard data export option, citing an access gap versus Free/Plus/Pro. A new GitHub tool (“scrapemychats”) provides an unofficial workaround to locally archive accessible chat logs (including attachments), with a 600-chat download taking “a few hours” due to throttling. While this is a user-data portability issue rather than a financial catalyst, it increases scrutiny around OpenAI’s data handling and potential compliance/terms-of-service risk.

Analysis

The market read-through is less about a direct monetization hit and more about enterprise trust. When data portability feels optional, procurement teams respond by adding exit clauses, archive requirements, and governance reviews, which tends to slow conversion for standalone AI tools over the next 1-3 months even if usage stays high.

The likely winners are suite vendors and control-plane names that can sell “safe AI” rather than raw model access: MSFT and, to a lesser extent, GOOGL. Security/data-governance budgets should also get a second-order bump as CIOs respond with DLP, retention, and eDiscovery spend; that favors PANW, ZS, and VRNS more than pure-play AI app names. The losers are vendors whose differentiation depends on frictionless lock-in, because in enterprise software lock-in often converts into procurement pushback and slower renewals rather than higher lifetime value.

Contrarian view: the consensus may underappreciate how quickly this can become a competitive issue. If customers perceive that one vendor makes exit harder than rivals, the bias is not “stickier revenue” but dual-vendor adoption and a preference for reversible workflows. Falsifiers are straightforward: an admin-grade export/retention fix from OpenAI in the next few weeks, or no deceleration in enterprise adoption/churn commentary over the next quarter.

Key Decisions for Investors

• No immediate trade in OpenAI-adjacent public names; wait for 1 quarter of renewal/churn data before expressing a view, because this is primarily a procurement/friction story rather than a revenue event.
• Buy MSFT on any AI-related pullback over the next 1-4 weeks as the cleanest public proxy for enterprise AI trust and governance; reward is modest but durable if buyers continue shifting toward suite-based deployments, and the thesis is falsified if enterprise AI commentary weakens.
• Initiate a small 1-3 month long basket in PANW and VRNS on weakness to express the likely budget shift toward DLP, archive, and data-governance tooling; stop if security spend does not inflect in upcoming CIO commentary.
• Pair trade: long MSFT / short C3.ai (AI) for 1-3 months to capture the preference for integrated, controllable AI over standalone apps when portability becomes part of the RFP; risk is a momentum-driven re-rating of smaller AI names regardless of fundamentals.