Escalation Classifier

The EscalationClassifier in deepeval is a categorical LLM-as-a-judge that determines whether your LLM application handed the user to a human, offered to, or handled the request itself.

The most effective datasets pair each escalation trigger (an angry customer, a legal threat, an out-of-scope request, a mention of self-harm) with the behavior expected from the assistant in that situation.

Labels

• escalated: the response hands the user over to a human agent or another escalation path.
• offered: the response offers to escalate or connect the user with a human but does not do so yet.
• not_escalated: the response handles the request itself with no mention of escalation.

Usage

from deepeval.classifiers import EscalationClassifier
from deepeval.test_case import LLMTestCase
from deepeval import evaluate

classifier = EscalationClassifier()

test_case = LLMTestCase(
    input="I'm going to sue you if this isn't fixed today.",
    actual_output="I understand. I'm connecting you with a member of our team now.",
    expected_labels={classifier.name: "escalated"},
)

evaluate(test_cases=[test_case], classifiers=[classifier])

There are FIVE optional parameters when creating a EscalationClassifier:

• [Optional] model: a string specifying which of OpenAI's GPT models to use, OR any custom LLM model of type DeepEvalBaseLLM. Defaulted to gpt-5.4.
• [Optional] include_reason: a boolean which when set to True, includes a reason for the chosen label. Defaulted to True.
• [Optional] allow_none: a boolean which when set to True, lets the classifier return no label when none of them fit (surfaced as label=None, with a reason). When False, the closest label is always chosen. Defaulted to False.
• [Optional] async_mode: a boolean which when set to True, enables concurrent execution within the classify() method. Defaulted to True.
• [Optional] classification_template: a subclass of ClassifierTemplate used to override the default prompts. Defaulted to ClassifierTemplate.

As a standalone

You can also run the EscalationClassifier on a single test case as a standalone, one-off execution. classify() returns the label and stores the result on the instance:

...

label = classifier.classify(test_case)
print(classifier.label, classifier.reason)

How Is It Calculated?

The EscalationClassifier is a one-shot LLM-as-a-judge: a single call to your evaluation model with the test case content and the labels above, returning the chosen label and a reason. The result is compared against the test case's expected_labels entry, if any, to decide pass or fail.