I currently have a shared Google Workspace drive with many folders and subfolders containing all kinds of files, sheets, photos, videos, PDFs, etc., that is, mostly unstructured data. I want to create an agent in Google Cloud that I can request for a specific project, and it will bring back everything it finds about it. What technologies and how would you design the workflow? Thanks!
Personally, I would start with Notebook LM (notebooklm.google.com) for a quick no-code solution. But it won’t solve for some of your unstructured data as it supports file types: PDF, .txt, Markdown, Audio (e.g. mp3), YouTube videos, Workspace docs.
Next, I would evaluate Agentspace and keep an eye on the product releases.
And of course there is Vertex AI search (aka AI Applications) for enterprise use cases. What is AI Applications? | Google Cloud