diff --git a/rasa/core/policies/flow_policy.py b/rasa/core/policies/flow_policy.py
index 31c6ca93521c..fcd282e2233e 100644
--- a/rasa/core/policies/flow_policy.py
+++ b/rasa/core/policies/flow_policy.py
@@ -56,7 +56,6 @@
     FlowsList,
     GenerateResponseFlowStep,
     IfFlowLink,
-    EntryPromptFlowStep,
     SlotRejection,
     StepThatCanStartAFlow,
     UserMessageStep,
@@ -677,10 +676,6 @@ def run_step(
             structlogger.debug("flow.step.run.branch")
             return ContinueFlowWithNextStep()
 
-        elif isinstance(step, EntryPromptFlowStep):
-            structlogger.debug("flow.step.run.entry_prompt")
-            return ContinueFlowWithNextStep()
-
         elif isinstance(step, GenerateResponseFlowStep):
             structlogger.debug("flow.step.run.generate_response")
             generated = step.generate(tracker)
diff --git a/rasa/shared/core/flows/flow.py b/rasa/shared/core/flows/flow.py
index 023c5f621349..656309c23564 100644
--- a/rasa/shared/core/flows/flow.py
+++ b/rasa/shared/core/flows/flow.py
@@ -639,8 +639,6 @@ def step_from_json(flow_step_config: Dict[Text, Any]) -> FlowStep:
         return LinkFlowStep.from_json(flow_step_config)
     if "set_slots" in flow_step_config:
         return SetSlotsFlowStep.from_json(flow_step_config)
-    if "entry_prompt" in flow_step_config:
-        return EntryPromptFlowStep.from_json(flow_step_config)
     if "generation_prompt" in flow_step_config:
         return GenerateResponseFlowStep.from_json(flow_step_config)
     else:
@@ -1144,111 +1142,6 @@ def default_id_postfix(self) -> str:
         return "generate"
 
 
-@dataclass
-class EntryPromptFlowStep(FlowStep, StepThatCanStartAFlow):
-    """Represents the configuration of a step prompting an LLM."""
-
-    entry_prompt: Text
-    """The prompt template of the flow step."""
-    advance_if: Optional[Text]
-    """The expected response to start the flow"""
-    llm_config: Optional[Dict[Text, Any]] = None
-    """The LLM configuration of the flow step."""
-
-    @classmethod
-    def from_json(cls, flow_step_config: Dict[Text, Any]) -> EntryPromptFlowStep:
-        """Used to read flow steps from parsed YAML.
-
-        Args:
-            flow_step_config: The parsed YAML as a dictionary.
-
-        Returns:
-            The parsed flow step.
-        """
-        base = super()._from_json(flow_step_config)
-        return EntryPromptFlowStep(
-            entry_prompt=flow_step_config.get("entry_prompt", ""),
-            advance_if=flow_step_config.get("advance_if"),
-            llm_config=flow_step_config.get("llm", None),
-            **base.__dict__,
-        )
-
-    def as_json(self) -> Dict[Text, Any]:
-        """Returns the flow step as a dictionary.
-
-        Returns:
-            The flow step as a dictionary.
-        """
-        dump = super().as_json()
-        dump["entry_prompt"] = self.entry_prompt
-        if self.advance_if:
-            dump["advance_if"] = self.advance_if
-
-        if self.llm_config:
-            dump["llm"] = self.llm_config
-        return dump
-
-    def _generate_using_llm(self, prompt: str) -> Optional[str]:
-        """Use LLM to generate a response.
-
-        Args:
-            prompt: the prompt to send to the LLM
-
-        Returns:
-            generated text
-        """
-        from rasa.shared.utils.llm import llm_factory
-
-        llm = llm_factory(self.llm_config, DEFAULT_LLM_CONFIG)
-
-        try:
-            return llm(prompt)
-        except Exception as e:
-            # unfortunately, langchain does not wrap LLM exceptions which means
-            # we have to catch all exceptions here
-            structlogger.error(
-                "flow.entry_step.llm.error", error=e, step=self.id, prompt=prompt
-            )
-            return None
-
-    def is_triggered(self, tracker: DialogueStateTracker) -> bool:
-        """Returns whether the flow step is triggered by the given intent and entities.
-
-        Args:
-            intent: The intent to check.
-            entities: The entities to check.
-
-        Returns:
-            Whether the flow step is triggered by the given intent and entities.
-        """
-        from rasa.shared.utils import llm
-        from jinja2 import Template
-
-        if not self.entry_prompt:
-            return False
-
-        context = {
-            "history": llm.tracker_as_readable_transcript(tracker, max_turns=5),
-            "latest_user_message": tracker.latest_message.text
-            if tracker.latest_message
-            else "",
-        }
-        context.update(tracker.current_slot_values())
-        prompt = Template(self.entry_prompt).render(context)
-
-        generated = self._generate_using_llm(prompt)
-
-        expected_response = self.advance_if.lower() if self.advance_if else "yes"
-        if generated and generated.lower() == expected_response:
-            return True
-        else:
-            return False
-
-    def default_id_postfix(self) -> str:
-        """Returns the default id postfix of the flow step."""
-        return "entry_prompt"
-
-
 @dataclass
 class SlotRejection:
     """A slot rejection."""
diff --git a/rasa/shared/core/flows/flows_yaml_schema.json b/rasa/shared/core/flows/flows_yaml_schema.json
index dd5b08b2a89f..3435b991310d 100644
--- a/rasa/shared/core/flows/flows_yaml_schema.json
+++ b/rasa/shared/core/flows/flows_yaml_schema.json
@@ -151,23 +151,6 @@
                 "$ref": "#$defs/next"
               }
             }
-          },
-          {
-            "required": [
-              "entry_prompt"
-            ],
-            "additionalProperties": false,
-            "properties": {
-              "id": {
-                "type": "string"
-              },
-              "entry_prompt": {
-                "type": "string"
-              },
-              "next": {
-                "$ref": "#$defs/next"
-              }
-            }
           }
         ]
       }