When to reach for it
- Data is received from unstructured or unreliable environments.
- Cognitive data quality is essential to prevent hallucinations downstream.
Integrator.
Sanitize the percepts before reasoning touches them.
A validation pattern — a dedicated gate that validates all incoming information (percepts) for quality and consistency before it ever reaches the model.
In practiceA support agent's CRM lookup starts returning a `status` field that changed from a fixed enum to free text; the integrator rejects the payload against its schema and raises a typed error, instead of letting the model reason over a value it would silently misread.
When it backfires
- Input data is already strictly sanitized and verified upstream.
The tradeoff
Higher data consistency is achieved against additional latency during perception processing.
What it actually does.
Incoming percepts pass a schema check; malformed ones take the error branch.
Two ways this pattern will hurt you.
Validation runs but its verdict is ignored
The integrator flags a payload as invalid and the rest of the graph processes it anyway because the flag was never wired into the downstream router.
Fix · Reify the verdict as a typed state field. Add a conditional edge that routes invalid payloads to an explicit error branch.
The gate hardens until it rejects legitimate input
Each malformed payload tightens the schema a little further, until a legitimate variant the upstream system has always sent is rejected too — and the run dead-ends with no path forward.
Fix · Version the schema against the upstream system's documented contract, not against the last bad payload you saw; log rejected payloads with their reason so an over-tight rule becomes visible.
Where Integrator is native.
AllcustomAdaptable
Keep going