diff --git a/cookbook/reasoning/README.md b/cookbook/reasoning/README.md
new file mode 100644
index 000000000..47ed031ce
--- /dev/null
+++ b/cookbook/reasoning/README.md
@@ -0,0 +1,44 @@
+# Agentic Reasoning
+
+> WARNING: Reasoning is an experimental feature and may not work as expected.
+
+### Create and activate a virtual environment
+
+```shell
+python3 -m venv ~/.venvs/aienv
+source ~/.venvs/aienv/bin/activate
+```
+
+### Install libraries
+
+```shell
+pip install -U openai phidata
+```
+
+### Export your `OPENAI_API_KEY`
+
+```shell
+export OPENAI_API_KEY=***
+```
+
+### Run a reasoning agent that DOES NOT WORK
+
+```shell
+python cookbook/reasoning/strawberry.py
+```
+
+### Run other examples of reasoning agents
+
+```shell
+python cookbook/reasoning/logical_puzzle.py
+```
+
+```shell
+python cookbook/reasoning/ethical_dilemma.py
+```
+
+### Run reasoning agent with tools
+
+```shell
+python cookbook/reasoning/finance_agent.py
+```
diff --git a/cookbook/reasoning/strawberry.py b/cookbook/reasoning/strawberry.py
index c491a483c..dc6db1acf 100644
--- a/cookbook/reasoning/strawberry.py
+++ b/cookbook/reasoning/strawberry.py
@@ -4,7 +4,7 @@
 from phi.model.openai import OpenAIChat
 from phi.cli.console import console
 
-task = "How many 'r' are in the word 'strawberry'? Count alphabet by alphabet."
+task = "How many 'r' are in the word 'strawberry'?"
 
 regular_agent = Agent(model=OpenAIChat(id="gpt-4o"), markdown=True)
 reasoning_agent = Agent(model=OpenAIChat(id="gpt-4o"), reasoning=True, markdown=True, structured_outputs=True)
@@ -13,15 +13,10 @@
 async def main():
     console.rule("[bold blue]Counting 'r's in 'strawberry'[/bold blue]")
 
-    console.rule("[bold green]Regular Agent Stream)[/bold green]")
+    console.rule("[bold green]Regular Agent[/bold green]")
     await regular_agent.aprint_response(task, stream=True)
-    console.rule("[bold yellow]Reasoning Agent Stream)[/bold yellow]")
+    console.rule("[bold yellow]Reasoning Agent[/bold yellow]")
     await reasoning_agent.aprint_response(task, stream=True)
 
-    console.rule("[bold green]Regular Agent)[/bold green]")
-    await regular_agent.aprint_response(task)
-    console.rule("[bold yellow]Reasoning Agent)[/bold yellow]")
-    await reasoning_agent.aprint_response(task)
-
 
 asyncio.run(main())
diff --git a/phi/agent/agent.py b/phi/agent/agent.py
index fa05bf072..a14789a28 100644
--- a/phi/agent/agent.py
+++ b/phi/agent/agent.py
@@ -1181,10 +1181,12 @@ def save_run_response_to_file(self, message: Optional[Union[str, List, Dict, Mes
     def get_reasoning_agent(self, model: Optional[Model] = None) -> Agent:
         return Agent(
             model=model,
-            description="You are a meticulous and thoughtful agent that solves problems by working through them step-by-step.",
+            description="You are a meticulous and thoughtful assistant that solves a problem by thinking through it step-by-step.",
             instructions=[
-                "First - Analyze the task by carefully examining the request and develop a step-by-step plan to solve it.",
-                "Then work through the plan step-by-step, executing any tools as needed. For each step, provide:\n"
+                "First - Carefully analyze the task by spelling it out loud.",
+                "Then, break down the problem by thinking through it step by step and develop multiple strategies to solve the problem."
+                "Then, examine the users intent develop a step by step plan to solve the problem.",
+                "Work through your plan step-by-step, executing any tools as needed. For each step, provide:\n"
                 "  1. Title: A clear, concise title that encapsulates the step's main focus or objective.\n"
                 "  2. Action: Describe the action you will take in the first person (e.g., 'I will...').\n"
                 "  3. Result: Execute the action by running any necessary tools or providing an answer. Summarize the outcome.\n"
@@ -1203,6 +1205,10 @@ def get_reasoning_agent(self, model: Optional[Model] = None) -> Agent:
                 "  - If next_action is continue, proceed to the next step in your analysis.\n"
                 "  - If next_action is validate, validate the result and provide the final answer.\n"
                 "  - If next_action is final_answer, stop reasoning.",
+                "Remember - If next_action is validate, you must validate your result\n"
+                "  - Ensure the answer resolves the original request.\n"
+                "  - Validate your result using any necessary tools or methods.\n"
+                "  - If there is another method to solve the task, use that to validate the result.\n"
                 "Ensure your analysis is:\n"
                 "  - Complete: Validate results and run all necessary tools.\n"
                 "  - Comprehensive: Consider multiple angles and potential outcomes.\n"
@@ -1213,7 +1219,7 @@ def get_reasoning_agent(self, model: Optional[Model] = None) -> Agent:
                 "  - Remember to run any tools you need to solve the problem.\n"
                 f"  - Take at least {self.reasoning_min_steps} steps to solve the problem.\n"
                 "  - If you have all the information you need, provide the final answer.\n"
-                "  - Remember to run any tools you need to run to solve the problem.",
+                "  - IMPORTANT: IF AT ANY TIME THE RESULT IS WRONG, RESET AND START OVER.",
             ],
             tools=self.tools,
             show_tool_calls=False,
diff --git a/pyproject.toml b/pyproject.toml
index 9aa8e916b..7d16fd087 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [project]
 name = "phidata"
-version = "2.5.8"
+version = "2.5.9"
 description = "Build AI Agents with memory, knowledge and tools."
 requires-python = ">=3.7"
 readme = "README.md"