Lesson: Foundations · Topic 2 of 2
Lesson summary
Bluejay is built around a small set of core entities that work together. Understanding how they connect is the key to using the platform effectively. This topic walks through each entity, explains what it does in plain language, and shows how they all fit together — like pieces of a puzzle.Objectives
- Map out the relationship between Agents, Simulations, Digital Humans, and Communities
- Understand what Custom Metrics are and why they matter
- Know where each entity lives in the dashboard
Video walkthrough
Video coming soon. Follow the written walkthrough below in the meantime.
Walkthrough
Meet the core entities
Here is a quick reference for the five main entities you will work with:
| Entity | What it is | Real-world analogy | Dashboard location |
|---|---|---|---|
| Agent | A mirror of your production AI agent | The product you are testing | Agents page |
| Simulation | A repeatable test that runs conversations against an Agent | A test case | Simulations page |
| Digital Human | A synthetic persona that plays the caller or customer | A test actor reading a script | Inside a Simulation |
| Community | A reusable group of Digital Humans | A casting pool of actors | Communities page |
| Custom Metric | An evaluation rule scored after each conversation | A grading rubric | Custom Metrics page |
See how they connect
The entities relate to each other like this:An Agent can have many Simulations. Each Simulation includes one or more Digital Humans — you can create them inside the simulation or pull them from a Community. When a conversation finishes, Custom Metrics score the result automatically.
Understand each entity's role
- Agent: The starting point. Everything you do in Bluejay — testing, monitoring, metrics — is organized under an Agent.
- Simulation: Think of it as a test case. You define the scenario, attach the actors (Digital Humans), pick the grading rubric (Custom Metrics), and press “Run.”
- Digital Human: The actor. It has a persona (personality, voice style) and a scenario script (what it is calling about). It is what makes your test feel like a real conversation.
- Community: A folder for Digital Humans you want to reuse. If you build a set of ten personas that cover your most common call types, put them in a Community so you can attach the whole group to any simulation with one click.
- Custom Metric: The grading rubric. It answers a specific question about the conversation, such as “Did the agent confirm the order number?” or “Was the greeting delivered in the first ten seconds?” Metrics are used in both testing and production monitoring.
Activity
Knowledge check
Can a Digital Human belong to multiple Communities?
Can a Digital Human belong to multiple Communities?
Yes. Digital Humans are standalone entities that can be added to any number of Communities, making it easy to reuse personas across different Simulations.
When are Custom Metrics evaluated?
When are Custom Metrics evaluated?
Custom Metrics are evaluated after a conversation ends — whether that conversation was a simulation run or a production call ingested through observability.
What is the difference between a Digital Human created inline and one pulled from a Community?
What is the difference between a Digital Human created inline and one pulled from a Community?
There is no structural difference — they are the same kind of entity. The only difference is organization: inline Digital Humans live inside a single Simulation, while Community members can be shared across many Simulations.