Sapience Objects & Resources

Understanding the different things is one thing, but the more powerful approach is to understand how they interconnect. A “graph” is just a diagram connecting things to other things (nodes) with relationships (”edge”). Here’s 1,000 words worth of a picture (click it for larger):

The core building block. All Users belong to an Org. Sapience is an enterprise software platform so its built around orgs. You can think of an Org as your company, e.g. “Acme”. But if you are a government department, you would use that. Also note that Orgs can tier, and have sub-orgs. So if a multi-national company is using the system (which they are!) its likely they would have one Org for HQ, and child Orgs under that for each of their business units.

By now you should know what this is 🙂 Agents that you deal with generally belong to an Org or are of Scope “global”, which means they’re available to all users. There are also user-scoped Agents, which are Agents that users with the “agent creation” permission turned on have created.

This is you, dear reader.

A Conversation is a series of Messages between Users (one or more) and Agents (one or more). The simplest Conversation is when you go to the Chat interface and interact with an Agent. This is a Conversation with 1 Agent and 1 User. Group chats where there are multiple Agents (Team of Agents we call this) is still a Conversation, its just that there are multiple Agents. Group chats with other users are also still Conversations, they just have multiple Users in them.

You know what a File is. Sapience brings lots of innovations to cloud file storage being AI-native, but at its heart a Sapience File is still a binary object… i.e. ones and zeros from a disk somewhere.

Sapience’s native document. Global Notes are published by the Sapience team. In your workspace you will likely have some global notes from us, org-scoped Notes that have been published by your Sapience Org Admin, and then the rest will be user-scoped Notes. i.e. Notes that you create yourself.

Task: a unit of work, assigned to a user (or an Agent), with a due date, and with notes and potentially with related resource links pointing to Files or Notes.

Goal: generally a phase of a project or used to group a series of Tasks together into a logical unit.

So, you can see from the diagram above, that the Project is different and interesting. It is both a container of all the other things, but it is also an Agent. This dual nature makes Projects crazy-powerful and the best part of Sapience.

Projects as workspace: because a Project can contain everything in Sapience, when you create a Project you are creating a little slice of the app, or better thought of as a workspace. Its excellent for organizing your work with Agents and with other humans.

Projects as Agents: because a Project is also an Agent, that has in its context all of the Goals, Tasks, Notes, Files and Project meta-data (ETA, description, success criteria, etc.) you have your own customized AI Agent, running on the latest AI model, with a beautifully curated set of context for it to work within. You can create “naked Agents” in the system, where you build everything up manually - but the Project Agent is far superior in real-life performance because of all the context it has. Context is King with AI systems.

There is actually a lot more than this going on behind the scenes, and in particular in support of the Org and Agent system, but this article is focused on helping Users understand Sapience. Developers, please refer to the API documentation.