---
title: "Best practices"
manual: "AI Assistant"
version: "main"
source: "UserGuide/BestPractices.rst"
rendered: "2026-10-03T14:54:45+00:00"
---

# Best practices {#best-practices}

## Start simple {#start-simple}

Begin with one assistant, one collection and a short pipeline. Add query
optimization, context optimization and quality gates only when you can observe a
specific need in traces.

## Use metadata deliberately {#use-metadata-deliberately}

Metadata is essential for source attribution and for answering questions about
where information comes from. Keep metadata stable and do not rely on generated
text for source identity.

## Keep prompts step-specific {#keep-prompts-step-specific}

Avoid putting every instruction into the assistant profile. Step-specific rules
make pipeline behaviour easier to test and debug.

## Test with ambiguous queries {#test-with-ambiguous-queries}

Use test questions with abbreviations, follow-ups, conflicting documents and
source questions. These queries reveal retrieval and prompt issues quickly.

## Avoid duplicate indexing {#avoid-duplicate-indexing}

When several indexers can write into the same collection and vector store, they
must agree on stable source identifiers. Otherwise the same content can appear
multiple times in retrieval.