Hey everyone, hoping to get some collective wisdom on a ContentBot hiccup I've hit this week. I've been deep in the process of setting up a new brand voice model for a client, aiming to really nail their specific tone—think "authoritative but approachable B2B SaaS."
I followed the brand voice training steps to the letter. I uploaded a solid batch of source documents: about 15 PDFs including past blog posts, whitepapers, and a handful of key website pages. The upload seemed to go fine, no error messages. The system said it was processing and training the model. I left it running overnight.
Here's the kicker: I came back in the morning, and the training was marked as "complete." But when I went to check the model or use it, there's nothing there. The model list is empty. Even weirder, when I go back to the "Documents" section for that brand voice project, it's also completely empty—like it never ingested anything. It's as if the training process just... ate my documents and produced no output.
Has anyone else run into this? I'm trying to troubleshoot and my immediate thoughts are:
* **Document format issues?** The PDFs were all text-based, not scanned images, so that *should* be okay.
* **File size or quantity?** The total was under 50MB, which I believe is within limits.
* **Processing timeout?** Could it have silently failed and just cleared the queue?
This is a bit of a blocker for automating our content briefs. Before I reach out to support, I wanted to see if the community has any practical steps.
* Did you need to "save" or "publish" the model somewhere non-obvious after training?
* Is there a dashboard or log to see if the training actually failed?
* Should I try again with just 2-3 documents first to test the pipeline?
Any shared experiences or workflow tips would be hugely appreciated. It's frustrating to lose that setup time and have to start from scratch without knowing the root cause.
- Al
Automate the boring stuff.