Chunking

Chunking is splitting documents into smaller pieces so they can be indexed and retrieved for AI systems.

Written by Reflective Data experts · updated Oct 5, 2026

Chunk size and boundaries strongly affect answer quality: too small loses context, too large dilutes relevance. Splitting along headings and keeping titles with each chunk usually works better than fixed sizes.

Test retrieval with real questions.

Embeddings Retrieval-augmented generation (RAG)Vector database

// how we can help

We build company brains, agent harnesses and AI bots on top of data you can trust. AI services → · The data platform behind it

ISO/IEC 42001

ISO/IEC 42001 is an international standard for an AI management system.

hallucination.md

Hallucination

A hallucination is when an AI model produces confident but incorrect or invented information.

system-prompt.md

System prompt

A system prompt is the standing set of instructions that defines how an AI assistant behaves…

fine-tuning.md

Fine-tuning

Fine-tuning is training an existing model further on your own examples, to specialise its behaviour or…

text-to-sql.md

Text-to-SQL

Text-to-SQL is using AI to turn a plain-language question into a SQL query.

ai-observability.md

AI observability

AI observability is tracking what an AI system does in production: inputs, outputs, tool calls, cost,…

← full analytics dictionary