Improve Subagent Classification with HyperClassifier

New agents created from the default Agentforce Service agent and Agentforce Employee agent templates in the new Agentforce Builder now use the Salesforce-owned HyperClassifier model for subagent classification. The HyperClassifier model outperforms standard LLMs in speed and precision, particularly when following complex instructions or adhering to negative instructions.

Note Beginning in April 2026, agent topics are now called subagents. There are no changes to functionality. During this transition, you may see a mix of the new and previous terms in our documentation.

Where: This change applies to Lightning Experience in Enterprise, Performance, Unlimited, and Developer editions with Foundations or Agentforce 1 editions.

When: This change is available starting the week of April 6, 2026.

How: New agents created from the default Agentforce Service agent and Agentforce Employee agent templates use HyperClassifier automatically so there's nothing for you to do.

Customize the model used in any subagent block by using the model_config property in Agent Script. You can't customize the model in the Agent Router in Canvas view.

This start_agent block in Agent Script configures HyperClassifier for subagent classification.

start_agent agent_router:
    description: "Welcome the user and determine the appropriate topic/subagent based on user input"
    model_config:
       model: "model://sfdc_ai__DefaultEinsteinHyperClassifier"

If you remove model_config, your agent defaults for subagent classification to the model option selection for Agentforce in Setup.

For more details, see Model Config Block in the Agentforce Developer Guide.