> For the complete documentation index, see [llms.txt](https://docs.algenta.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.algenta.ai/guides/decision-memory.md).

# Log decisions & record outcomes

By the end of this guide you will have logged a decision to decision memory, recorded the real number once you knew it, and read back the engine-computed `outcome_delta` that closes the loop. You need an Algenta API key (see [Authentication](/getting-started/authentication.md)) exported as `$ALGENTA_API_KEY`, and either `curl`, the `de` CLI, or the Python SDK (`pip install algenta-sdk`). The loop is `log -> outcome -> review`; only the first step is required, and the outcome can be recorded minutes or weeks later.

{% stepper %}
{% step %}

### Log the decision

Send the action you chose to `POST /v1/decisions`. Only `chosen_action` is required; logging `expected_value` is what makes the engine able to compute a delta later, and an optional `run_id` binds the record to the simulation that produced it. The call returns `201 Created` with the new decision `id` — capture it.

{% tabs %}
{% tab title="HTTP" %}

```bash
DECISION_ID=$(curl -s -X POST https://api.algenta.ai/v1/decisions \
  -H "Authorization: Bearer $ALGENTA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
        "context": "Q3 pricing change for the Pro tier",
        "chosen_action": "raise_price_10pct",
        "options_considered": ["hold", "raise_price_10pct", "raise_price_20pct"],
        "expected_value": 142000,
        "confidence": 0.78,
        "risk_p5": -8000,
        "risk_p95": 210000,
        "risk_pol": 0.12
      }' | jq -r .id)
echo "$DECISION_ID"
```

{% endtab %}

{% tab title="CLI" %}

```bash
# Put the request body in a JSON file, then log it.
de decisions log decision.json
```

{% endtab %}

{% tab title="Python" %}

```python
import os
from decision_engine import AlgentaClient

client = AlgentaClient(api_key=os.environ["ALGENTA_API_KEY"])

decision = client.log_decision(
    {
        "context": "Q3 pricing change for the Pro tier",
        "chosen_action": "raise_price_10pct",
        "options_considered": ["hold", "raise_price_10pct", "raise_price_20pct"],
        "expected_value": 142000,
        "confidence": 0.78,
        "risk_p5": -8000,
        "risk_p95": 210000,
        "risk_pol": 0.12,
    }
)
print(decision.id)
```

{% endtab %}
{% endtabs %}

Verify by confirming the response carried a non-empty `id` and a `created_at` timestamp. The record also pins `policy_snapshot_id`, `schema_snapshot_id`, and `manifest_version`, which is what makes the decision auditable later.
{% endstep %}

{% step %}

### Record the actual outcome

Once the real number is known, close the loop with `PATCH /v1/decisions/{id}/outcome`. The body requires `actual_outcome` (a float) and accepts optional `outcome_notes`. The engine sets `outcome_recorded_at` and computes `outcome_delta = actual_outcome - expected_value` whenever an `expected_value` was logged. This call also refreshes your org's confidence calibration.

{% tabs %}
{% tab title="HTTP" %}

```bash
curl -s -X PATCH "https://api.algenta.ai/v1/decisions/$DECISION_ID/outcome" \
  -H "Authorization: Bearer $ALGENTA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"actual_outcome": 131500, "outcome_notes": "Lower conversion than modeled."}'
```

{% endtab %}

{% tab title="CLI" %}

```bash
# outcome.json holds {"actual_outcome": 131500, "outcome_notes": "..."}
de decisions outcome "$DECISION_ID" outcome.json
```

{% endtab %}

{% tab title="Python" %}

```python
updated = client.record_outcome(
    decision.id,
    actual_outcome=131500,
    outcome_notes="Lower conversion than modeled.",
)
print(updated.outcome_delta)  # -10500
```

{% endtab %}
{% endtabs %}

Verify by checking the `200 OK` response carries `outcome_delta` (here `131500 - 142000 = -10500`) and a populated `outcome_recorded_at`. A negative delta means the result came in worse than the engine expected.
{% endstep %}

{% step %}

### Review the closed loop

Re-read the persisted record to confirm the loop is closed, and list decisions to compare expected against actual across the org. `GET /v1/decisions/{id}` returns the single record; `GET /v1/decisions?with_outcome_only=true` returns only decisions that already have a recorded outcome, most recent first.

{% tabs %}
{% tab title="HTTP" %}

```bash
# Re-read the single decision
curl -s "https://api.algenta.ai/v1/decisions/$DECISION_ID" \
  -H "Authorization: Bearer $ALGENTA_API_KEY" | jq '{id, expected_value, actual_outcome, outcome_delta}'

# List only decisions that have an outcome recorded
curl -s "https://api.algenta.ai/v1/decisions?with_outcome_only=true&limit=20" \
  -H "Authorization: Bearer $ALGENTA_API_KEY" | jq '.total'
```

{% endtab %}

{% tab title="CLI" %}

```bash
de decisions get "$DECISION_ID"
de decisions list --with-outcome-only --limit 20
```

{% endtab %}

{% tab title="Python" %}

```python
record = client.get_decision(decision.id)
print(record.expected_value, record.actual_outcome, record.outcome_delta)

reviewed = client.list_decisions(with_outcome_only=True, limit=20)
print(reviewed.total)
```

{% endtab %}
{% endtabs %}

Verify by confirming the single record shows your `actual_outcome` and `outcome_delta`, and that the filtered list `total` includes the decision you just closed.
{% endstep %}
{% endstepper %}

## Expected result

You logged one decision (HTTP `201`), recorded its outcome (HTTP `200`), and read back a closed-loop record. The single decision now carries `expected_value: 142000`, `actual_outcome: 131500`, and an engine-computed `outcome_delta: -10500`, alongside `outcome_recorded_at` and the integrity snapshot fields.

```json
{
  "id": "…",
  "chosen_action": "raise_price_10pct",
  "expected_value": 142000,
  "actual_outcome": 131500,
  "outcome_delta": -10500,
  "outcome_recorded_at": "2026-06-21T13:30:00Z",
  "policy_snapshot_id": "…",
  "schema_snapshot_id": "…",
  "manifest_version": "…"
}
```

## Troubleshooting

{% hint style="warning" %}
**`outcome_delta` is `null` after recording an outcome.** The delta is only computed when an `expected_value` was present on the original log. Log the decision with `expected_value`, or read the raw `actual_outcome` instead.
{% endhint %}

{% hint style="info" %}
**`404 Decision not found` on the PATCH or GET.** Decision memory is scoped to your org and workspace. Confirm the `id` came from a `POST /v1/decisions` made with the same API key, and that the record was not soft-deleted with `DELETE /v1/decisions/{id}`. See [API errors](/http-api/errors.md).
{% endhint %}

## Next

{% content-ref url="/pages/ktXIojcv61EsP0IhlJTb" %}
[Decision memory](/concepts/decision-memory.md)
{% endcontent-ref %}

{% content-ref url="/pages/dbMexAzWYilfd77tn3O8" %}
[API endpoint reference](/http-api/reference.md)
{% endcontent-ref %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.algenta.ai/guides/decision-memory.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.