Skip to main content
Digital Humans are synthetic replicas of your customers. They are highly configurable test entities that interact with your agent during Bluejay simulations, enabling you to validate behavior, catch edge cases, and build confidence before going live. Think of Digital Humans as the test cases of your agentic test suite. Just like unit tests exercise individual functions, Digital Humans exercise individual conversation scenarios against your agent. The best way to create them is to start by thinking about how your real customers will interact with your agent.

​
What You’ll Learn

  • What a Digital Human represents in Bluejay
  • The anatomy of a Digital Human: intent, success criteria, and behavior
  • How Digital Humans drive realistic simulation conversations
  • Why Digital Humans are central to every evaluation on Bluejay

​
The Mental Model

Every simulation run pairs one or more Digital Humans against your agent. Each Digital Human carries a specific intent (what it wants to talk about) and success criteria (how Bluejay determines whether the interaction went well). After the conversation, Bluejay evaluates the transcript against those criteria and any custom metrics you’ve attached.

​
Anatomy of a Digital Human

A Digital Human is composed of three core layers:
1

Intent

The intent defines what the Digital Human will talk about during the conversation. This might be a billing dispute, a product return, an appointment booking, or any scenario your agent should handle. The more specific the intent, the more targeted the test.
2

Success Criteria

Success criteria define how Bluejay determines whether the interaction was successful. These are the conditions that must be met for the test to pass — for example, “the agent offered a refund” or “the agent correctly transferred the call to a supervisor.”
3

Behavior & Traits

Behavior settings control how the Digital Human sounds and acts during the conversation. This includes accent, language, emotion, speaking speed, background noise, and more. These traits make every Digital Human feel like a distinct, realistic customer.

​
Why Digital Humans Matter

Digital Humans are one of the most important concepts on Bluejay. They are the individual test case that will run against your agent. Without well-crafted Digital Humans, your simulations won’t reflect reality.

Realistic coverage

Each Digital Human represents a real-world customer scenario. Build a diverse set and you’ll cover the full spectrum of interactions your agent will face.

Repeatable benchmarks

Group Digital Humans into Communities to reuse the same benchmark population across agents, versions, and evaluation cycles.

Automated at scale

Generate up to 100 diverse Digital Humans in a single API call. The more you tell Bluejay about your agent’s behavior, the better we can generate a diverse range of Digital Humans.

Configurable behaviors

Fine-tune accents, languages, emotions, DTMF tones, scripted responses, IVR navigation, silence, and more to match any real-world scenario.

​
How Digital Humans Fit into the Workflow

1

Describe your agent

Tell Bluejay what your agent does, who it serves, and what success looks like. This context drives Digital Human generation.
2

Create or generate Digital Humans

Build individual Digital Humans manually, or let Bluejay generate a diverse population based on your agent’s description and goals.
3

Run simulations

Launch a simulation and each Digital Human calls or chats with your agent independently, producing a transcript for every interaction.
4

Evaluate results

Bluejay evaluates each transcript against the Digital Human’s success criteria and your custom metrics, giving you a pass/fail result with detailed scoring.

​
Next Steps

Configuration

Learn how to configure accents, DTMF, scripted responses, IVR simulation, and more.

Use Cases

See real-world scenarios and patterns for building effective Digital Humans.

Deep Dive

Full reference for Digital Human fields, generation, and best practices.