Data Sources Page

September 1, 2023

The Data Sources page is linked to in the sidebar on the platform. This is where you can take steps to take Humanitas AI to school and teach it about your organization and work. On this page you’ll find various Data Cards. You might think of each Data Card as a “class”. By uploading data and clicking into that card you can upload various documents to teach the AI. It’s important that the AI will only be able to learn from what you teach it i.e. what you’re able to upload.

Data Cards

Snapshot of the Datasources page within the partner.humanitas.ai portal

Each data card will correspond to a category of types of documents that Humanitas AI has been engineered to be able to process. One of our core values at Humanitas, as captured in our Data Charter, is around transparency. As such, each data card contains an explanation of what documents and data on what’s been shared (or what’s available from public sources the Humanitas team has processed) in that category. The card also includes an explanation of what actions that data enables Humanitas to take to help you through all of our AI and tools including but not limited to the AI Assistant.

Data Source Types

We’re always working to train Humanitas to learn to process new and powerful forms of data. As such, periodically new data sources will become available. You can always visit Feature Requests to recommend data types you would like to see Humanitas AI learn to process. You can also make requests there for improvements when you notice the tool is not effectively processing a data source. The data source types as of August 2023 include:

  • Other data sources (newsletters, blogs, grant applications, program descriptions, impact reports, pitch decks, letter templates)
  • 990s - with 990's you'll be invited to connect the 990s the system has on file for the nonprofit your are currently logged in as
  • Annual Reports

Note that you are limited to only three files at a time within the other category of data sources. This better allows the AI to process information and recognize recency and relevancy.