Synthetic data

Synthetic data is artificially generated data that mimics the properties of real data.

Written by Reflective Data experts · updated Oct 5, 2026

It is used to test pipelines and models without exposing personal information, and sometimes to augment small datasets. It must be validated, since poorly generated data can mislead.

It is not automatically free of privacy risk.

Data quality Large language model (LLM)Personally identifiable information (PII)

// 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

AI observability

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

function-calling.md

Function calling (tool use)

Function calling lets a language model request that your code run a defined tool, such as…

ai-evals.md

AI evaluations (evals)

AI evaluations are tests that measure how well an AI system performs on realistic tasks.

agent-harness.md

Agent harness

An agent harness is everything around a model that makes an agent useful and safe: tools,…

model-drift.md

Model drift

Model drift is when an AI or machine learning system gets worse over time because the…

context-window.md

Context window

The context window is the amount of text a language model can consider at once, including…

← full analytics dictionary