Data Sources Overview
AI works best when it has context, and the best context oftentimes comes from your business data. Using your workplace data in your Spells is simple with Data Sources.
What is Data Sources?
You can think of Data Sources as an AI-optimized searchable file repository. It’s a place for you to store and sync documents from your other systems, and each document in Data Sources is optimized for searching based on “meaning” (rather than keywords).
When you store a file in Data Sources, you’re able to search for the results that match the meaning of the search query. For example, let’s say you sync your Google Drive folder containing your product FAQs. If a user searches “How do I register an account?” it will look for text in your files related to registering an account, even if it’s called “signing up” or “free trials”. This is called “semantic search”, compared to “keyword search” where the query is matched against specific keywords in the data.
Semantic search is well-suited for AI use-cases, since AI often works with “unstructured” text - that is, text that doesn’t have a defined format or known context. For example, blog posts, emails, and research reports are unstructured; we don’t know their format or what words/content they may contain. This can make them difficult to search against with traditional methods, but semantic search usually works quite well.
Data Sources Overview
In the middle of the page, you’ll find your file store. The file store contains all of your files, which you can organize into folders.
At the top, you’ll find a search bar. Entering a query here will search your entire Data Sources repository for semantic matches. You can sort the results using the sort dropdown to the right of the search bar.
You can add files to Data Sources using the “New” button in the top right. You can choose to upload files from your computer, or sync files from different providers. We currently support:
- Google Drive
- Notion