Deep Dive Tool Error — What It Means & How to Fix It

What Is Deep Dive?

Deep Dive is a capability within Assist 2.0 that reads the full content of a file in your knowledge or attachments — rather than relying on the short summary shown in the system. Assist uses it automatically when it needs verbatim text, specific passages, or precise details from a document, image, audio, or video file to answer your request.

You won't usually trigger Deep Dive yourself. The agent decides when to use it based on what you've asked for (e.g. "quote the third paragraph," "summarise this PDF in detail," "what does the speaker say at 4:32").

What Does the Deep Dive Error Mean?

Sometimes Assist will report that Deep Dive was unable to process a file. This means the agent attempted to read the file in detail but couldn't complete the task. The rest of your response may still come through, but it won't be grounded in that file's contents.

Why Does This Error Happen?

The Deep Dive error can occur for a few reasons:

Unsupported file type — Deep Dive only supports a specific set of formats (see below). Anything outside that list will be silently skipped or error out.
Text or URL knowledge passed to Deep Dive — Plain text and URL-type knowledge aren't supported by Deep Dive. These should be read directly instead, and Assist will normally handle that automatically.
File is too large — The file exceeds the size limit the tool can process in one pass.
File is empty or unreadable — The file looks valid but contains no extractable content (e.g. a scanned PDF without OCR, a silent audio file).
Content is too complex or lengthy — Very long or dense files can cause the tool to time out.
Temporary system issue — Occasionally a short-lived processing issue causes the error.

Supported File Types

Deep Dive works with:

Documents — PDF, TXT, MD, EML
Images — PNG, JPG, JPEG, WEBP
Video — MP4, MOV, WEBM, and other common video formats
Audio — MP3, M4A, OGG, WAV, and other common audio formats

For text-type or URL-type knowledge, Deep Dive isn't used at all — the agent reads those directly via its knowledge-reading tool. If you've added a .docx , .csv , or .xlsx and want Assist to read it in detail, convert it to PDF or TXT first.

How to Fix It

1. Check Your File Format

Confirm the file is one of the supported types above. If it's a .docx , .csv , .xlsx , or anything else not on the list, convert it to PDF or TXT before re-uploading.

[SCREENSHOT: Knowledge upload panel in Assist 2.0 showing a file being added]

2. Check File Content

Make sure the file actually contains readable content:

Scanned PDFs need OCR before they can be read
Image files should be clear and legible
Audio and video files should have an audible track or visible content

3. Reduce File Size

If the file is very large, try:

Splitting it into smaller sections
Removing unnecessary pages, slides, or footage
Compressing audio/video to a lower bitrate

4. Re-upload the File

Occasionally the upload itself is the problem. Remove the file from your knowledge and add it again.

5. Try Again Shortly

If the error feels intermittent, wait a moment and re-run the request. Temporary processing issues usually resolve quickly.

6. Be Specific in Your Prompt

Deep Dive runs when the agent decides it's needed. If you want the agent to read a specific file in detail, say so directly — e.g. "read the attached PDF carefully and quote the section on pricing." This makes it more likely the agent will invoke Deep Dive on the right file.

Still Seeing the Error?

If you've tried the above and the error persists, it may be a known limitation or a bug. Contact your workspace admin or the Assist support team — include the file type, file size, the prompt you used, and a screenshot of the error if possible.