An update to my previous post: this was not a new issue first reported on October 29. I had already been corresponding with Support and providing the evidence they requested before then.
The following are the two issues I reported and supplied evidence for, only to be left waiting without a date for resolution or the next update, receive no reply to my follow-up, and then receive an answer about a different problem under a different case number when I contacted Support again:
- Codex Cloud failing to connect to GitHub.
- dot being unable to read or continue an existing LOCAL Codex task.
Here is the timeline. All times are Japan Standard Time (JST, UTC+9).
-
Before October 18: I was already corresponding with Support and had provided explanations and screenshots of the problems.
-
October 18, 16:23: Support requested a screen recording showing the existing failure record. I sent the requested recording at 16:43 the same day.
-
October 26, 18:55: Having received no substantive response after submitting the recording, I followed up to ask about the investigation, the next action, and a refund or other compensation.
-
October 29: With that follow-up still unanswered, I contacted Support again with a consolidated account of the situation. I was told that the request had been escalated to a specialist and that I could expect a response in the coming days. However, there was no confirmation that a specialist had reviewed the evidence already supplied, and no specific date for a response.
-
October 30, 01:02: I received a support email about a Windows Library connector issue. It did not address either of the two problems above. It also carried a different case number, without explaining which of my inquiries it related to.
-
October 30, 09:13: I replied to point out the mismatch and requested clarification of the case linkage, review of the existing evidence, answers to the two original issues, a date for the next update, and a response regarding a refund or credit.
-
October 30, 15:38: I received a reply to my morning email, signed “OpenAI AI Support Agent”. It advised me to open the existing Codex task and enter instructions directly, or select the task in dot’s Activity and respond to any approval prompt shown. It did not explain why dot’s read/continue workflow fails or how to resolve it. There was also no answer about the cloud connection failure, the results of reviewing my evidence, a refund, the responsible team, or the next update date.
I am not asking how to manually operate each task myself. I am reporting that the workflow for progressing a project through dot is not working. Instructions to perform the work manually do not resolve that report.
Frankly, I feel that a refund alone would not make up for this. I would also like compensation for the time spent and the additional checking and explanation this process has required. As a paying customer, I have supplied evidence, repeatedly reconstructed the history, and corrected mismatched responses, yet my core questions remain unanswered. The burden goes beyond the subscription fee.
What concerns me further is that other users’ public accounts describe similar experiences: being sent back to basic questions after submitting evidence, receiving replies that leave the actual questions unanswered, and being unable to establish whether a human specialist is investigating. These reports do not establish how widespread the problem is or whether it is intentional, but they make it difficult for me to dismiss my experience as an isolated misunderstanding.
If OpenAI’s leadership, including Sam Altman, wants investors and the public to believe in its vision for AI, it should first demonstrate accountability to the customers paying for its products today. The gap between the vision of AI benefiting humanity and the reality of customers being worn down by unanswered support requests is stark. From my experience, promotion and expansion appear to take priority over responsibility to existing customers. This has damaged my trust not only in the product, but also in whether OpenAI is acting as a responsible AI company.
I am asking for a specialist to review the evidence already supplied, address each issue specifically, explain the next action and update date, and provide a clear decision on a refund or other compensation. If an answer is not yet available, please explain what remains unverified and what will happen next.
At this point, I can’t help expecting that whatever I say will once again be ignored, passed from one support agent to another, or routed back to AI support, with the actual issues left unresolved. Apparently, the “Open” in OpenAI does not extend to its customers.