Unable to summarize session using new Teams

General
, ,
1

hi team,

I’ve added persistent storage to my team object. I want to add user memory and session summary as well along with session data. But I’m not able to do so. I’m able to save session though but not seeing user memory or session summary.

db_url = "postgresql+psycopg2://Shekhar@localhost:5432/agno"

session_id="9"
storage = PostgresStorage(
    table_name="sessions", schema="session", db_url=db_url
)
memory = TeamMemory(
    db=PgMemoryDb(
        table_name="agent_team_memory",
        schema="sessions",
        db_url=db_url,
    ),
    user_id="1",
    create_user_memories=True,
    update_user_memories_after_run=True,
    classifier=MemoryClassifier(model=AzureOpenAI(id="gpt-4o"))
)

loan_upselling_team:Team = Team(
    name="Loan Upselling team.",
    user_id="1",
    mode="coordinate",
    session_id=session_id,
    storage=storage,
    memory=memory,
    model=AzureOpenAI(id="gpt-4o"),
    members=[sales_agent, verification_agent, decision_making_agent],
    description="A team of agents handling customer queries and loan processes at Horizon Financial.",
    instructions=[
        "This is a team of specialized agents at Horizon Financial.",
        "Please route the user query to the agent which you think is responsible to handel that particular query"
        "The Sales Assistant leads customer interactions and delegates tasks.",
        "Collaborate to provide accurate, timely responses to customer queries.",
        "Other agent can ask you to gather customer input on their behalf that they need to provide response to customer query.",
        "After taking user input, please provide that user input back to that agent which has asked you to gather that from the user."
    ],
    success_criteria="""
    Following is the success criteria:
    - customer is satisfied with the response and his query is resolved.
    - customer has no further query to ask.
    - customer no longer wishes to continue the conversation..
    - customer is no longer interested in the loan offers.
    - customer has filled the loan application form and will be updated later on about the application status.
    """,
    add_context=True,
    add_state_in_messages=True,
    add_datetime_to_instructions=True,
    enable_agentic_context = True,
    share_member_interactions= True,
    read_team_history = True,
    enable_team_history=True,
    show_tool_calls=True,
    # # markdown=True,
    show_members_responses=True,
    # tools = [UserInputToolkit()],
    debug_mode=True,
)

Please assist @Monali @monalisha @kausmos @anuragphi

2

@Monali @monalisha @kausmos @anuragphi I’m blocked at the moment. Could you guys please guide me here ? I’ll really appreciate your prompt response. :folded_hands:

3

Hey @MahorShekhar
thanks for reaching out and supporting Agno. I’ve shared this with the team, we’re working through all requests one by one and will get back to you soon.
We appreciate your patience!

4

hey @Monali I really appreciate your continued assistance to me. Can I request you to please prioritize this. I’m really blocked at the moment.

5

Hey @MahorShekhar,
Our team is currently heads down, working toward an upcoming launch — but we’re doing our best to prioritize this and really appreciate your patience in the meantime :folded_hands:

Also, super curious — would love to hear more about what you’re building with Agno. Please share

6

@Monali I’m still stuck here.

Could you guys please guide me how to summarize team session using AzureOpenAI model. Also I believe you need to supply Memory Manage, Classifier, Summarizer explicitly since I’m using AzureOpenAI LLM model not OpenAI.

7

Hi @MahorShekhar
Apologies for not getting clarity here. We have just launched a new implementation of Memory that I think would be a lot cleaner to use for you.

Here is a link for memory in teams → Introduction - Agno

Here is the reference of how the classes work Memory - Agno

Here is how to use Memory in general Introduction - Agno

8

As to what we are building, we are currently looking at more sophisticated Human-in-the-loop features.

9

thanks @Dirk I saw those and already using them.

I would like to appreciate such good work from you and your team.