Python: [Python] Exception occurs when using Sequential Workflow with AzureAIClient

[q33566]

Environment

• agent-framework==1.0.0b251216
• agent-framework-azure-ai==1.0.0b251216
• agent-framework==1.0.0b251209
• agent-framework-azure-ai==1.0.0b251209

Description

I created two agents using AzureAIClient and added them to a sequential workflow.
When I run the workflow, an error occurs.

This is unexpected because the same setup worked correctly last week using version 1.0.0b251209.
However, this week it crashes even when using the same version.

Error Message

'agent_framework_azure_ai._client.AzureAIClient'> service failed to complete the prompt: Error code: 400 - 
{'error': 
    {'code': 'invalid_payload', 'message': 'Invalid payload', 'param': None, 'type': 'invalid_request_error', 'details': 
        [
            {'code': 'ValidationError', 'message': 'type: Value is "array" but should be "string"', 'param': '/input', 'type': 'error'},
            {'code': 'ValidationError', 'message': 'required: Required properties ["type"] are not present', 'param': '/input/1', 'type': 'error'},
            {'code': 'ValidationError', 'message': 'type: Value is "array" but should be "string"', 'param': '/input/1/content', 'type': 'error'}, 
            {'code': 'ValidationError', 'message': 'required: Required properties ["annotations"] are not present', 'param': '/input/1/content/0', 'type': 'error'}
        ], 
        'additionalInfo': {'request_id': '0ee672893116194c6d56a003b9ca6a8a'}
    }
}

Reproduce

import os
from typing import Any

from agent_framework import (
    ChatMessage,
    Role,
    SequentialBuilder,
)
from agent_framework.azure import AzureAIClient
from azure.ai.projects.aio import AIProjectClient
from azure.identity.aio import DefaultAzureCredential

project_client = AIProjectClient(
    endpoint=os.environ["AZURE_AI_PROJECT_ENDPOINT"],
    credential=DefaultAzureCredential(),
)
credential = DefaultAzureCredential()
client1 = AzureAIClient(
    project_client=project_client,
    async_credential=credential,
)
client2 = AzureAIClient(
    project_client=project_client,
    async_acredential=credential,
)
agent1 = client1.create_agent(name="agent1", instructions="Help user to plan tasks")
agent2 = client1.create_agent(name="agent2", instructions="refine the givenplan")
workflow = SequentialBuilder().participants([agent1, agent2]).build()

events = await workflow.run("Good morning! Help me plan my day.")
outputs = events.get_outputs()

if outputs:
    print("===== Final Conversation =====")
    messages: list[ChatMessage] | Any = outputs[0]
    for i, msg in enumerate(messages, start=1):
        name = msg.author_name or (
            "assistant" if msg.role == Role.ASSISTANT else "user"
        )
        print(f"{'-' * 60}\n{i:02d} [{name}]\n{msg.text}")

Related Issues

These issues all use AzureAIClient, which relies on the Azure AI Projects client.
The request format prepared by agent-framework is rejected by the Azure AI Projects
data plane REST APIs,
specifically the
create responses API,
used by the
Azure AI Project SDK.

• Python: AzureAIClient fails within GroupChatBuilder #2976
• Python: Multi‑turn with AgentFrameworkAgent + AzureAIClient to Azure AI Projects Foundry v2 agent fails on 2nd turn: agent_framework_azure_ai sends invalid input to agent/Responses #2964
• Python: [Python] AzureAIClient 400 error in workflow #3003

Root Cause

This issue may be caused by an input schema mismatch between the official OpenAI client and the OpenAI client obtained via AIProjectClient.get_openai_client() in Azure AI Projects.

agent-framework generates a structured input payload that is accepted by the official OpenAI Responses API, but the same payload is rejected by the OpenAI client created via the Azure AI Projects SDK (using the create responses API) with validation errors.

For more information, see: This issue in Azure AI Project SDK

[moonbox3]

Thanks for filing, @q33566. The fix (more a workaround, until the Azure SDK fixes their bug upstream) is in #3076.

When I run your sample, I get an output:

from typing import Any

from agent_framework import (
    ChatMessage,
    Role,
    SequentialBuilder,
)
from agent_framework.azure import AzureAIClient
from azure.identity.aio import DefaultAzureCredential


async def main():
    async with (
        DefaultAzureCredential() as credential,
        AzureAIClient(credential=credential).create_agent(
            name="agent1",
            instructions="Help user to plan tasks",
        ) as agent1,
        AzureAIClient(credential=credential).create_agent(
            name="agent2",
            instructions="Refine the given plan",
        ) as agent2,
    ):
        workflow = SequentialBuilder().participants([agent1, agent2]).build()
        events = await workflow.run("Good morning! Help me plan my day.")
        outputs = events.get_outputs()

        if outputs:
            print("===== Final Conversation =====")
            messages: list[ChatMessage] | Any = outputs[0]
            for i, msg in enumerate(messages, start=1):
                name = msg.author_name or ("assistant" if msg.role == Role.ASSISTANT else "user")
                print(f"{'-' * 60}\n{i:02d} [{name}]\n{msg.text}")


if __name__ == "__main__":
    import asyncio

    asyncio.run(main())

Output:

===== Final Conversation =====
------------------------------------------------------------
01 [user]
Good morning! Help me plan my day.
------------------------------------------------------------
02 [agent1]
Good morning! I’d be happy to help you plan your day. Could you tell me a bit about what you need to accomplish today and any appointments or commitments you have? Also, let me know if you prefer a detailed hourly schedule or just a general task list.
------------------------------------------------------------
03 [agent2]
Good morning! I’d be happy to help you plan your day. To get started, could you please share:

- Any appointments or meetings you have scheduled
- Tasks or projects you want to accomplish
- Personal activities or breaks you want to include
- Your preferred wake-up and sleep times
- Any specific priorities or deadlines

With that info, I can create a tailored plan to help you have a productive and balanced day!