Institutional learning

A research team that does not learn re-derives the same conclusions and repeats the same mistakes. The value compounds when an insight from one review is available to every agent on the next. Agno’s LearningMachine with a shared store and a global namespace makes the team’s learning institutional, not per-agent.

from agno.agent import Agent
from agno.learn import LearnedKnowledgeConfig, LearningMachine, LearningMode
from agno.models.openai import OpenAIResponses

# Shared learnings store, imported once, used by every agent and team
from agents.settings import team_learnings

analyst = Agent(
    name="Financial Analyst",
    model=OpenAIResponses(id="gpt-5.5"),
    learning=LearningMachine(
        knowledge=team_learnings,
        learned_knowledge=LearnedKnowledgeConfig(
            mode=LearningMode.AGENTIC,
            namespace="global",
        ),
    ),
)

# Another agent built on the same team_learnings store captures an insight:
risk_officer.print_response("Estimates lag this sector by about a quarter.")

# Later, a different agent reads it back from the shared store:
analyst.learning_machine.learned_knowledge_store.print(query="estimate lag")

Every analyst and every team writes to and reads from the same team_learnings store, so the capture above is already in the next agent’s context.

Per-agent vs institutional

	Per-agent memory	Institutional learning
Scope	One agent	Every agent and team that shares the store
Namespace	The user or agent	`"global"`
Effect	This analyst remembers	The committee remembers

For deep research, institutional is the point. The team’s judgment should outlive any single review.

What to capture

Worth a learning	Not worth a learning
”Analyst estimates lag this sector by a quarter”	A one-off number from a single query
”This data source double-counts renewals”	A restatement of the mandate
A correction to a conclusion that was wrong	A summary of what was already in the library

Capture corrections and transferable insights. Leave durable facts to the research library and rules to the static context.

Modes

Mode	Behavior	Use for
`ALWAYS`	Capture runs after every response	Steady accumulation of observations
`AGENTIC`	The agent decides what is worth keeping	Research, where signal-to-noise matters
`PROPOSE`	A human approves before it persists	Anything that changes how the team decides

This is the same machine the data agent uses to self-correct, pointed at a shared store instead of a per-warehouse one.

Next steps

Task	Guide
Feed it grounded context too	Grounding research
Audit what changed the team’s mind	Structured deliverable

Welcome

Get Started

Features

Use Cases

Institutional learning

Per-agent vs institutional

What to capture

Modes

Next steps

Developer Resources

​Per-agent vs institutional

​What to capture

​Modes

​Next steps

​Developer Resources

Per-agent vs institutional

What to capture

Modes

Next steps

Developer Resources