Daniel & Friends / Glossary / Grounding
AI and automationGrounding
Grounding means anchoring a model’s answer in verified sources, such as company documents, databases or web search results. The model draws on the supplied material, and the answer can be backed by a link. It reduces the risk of hallucinations but does not remove it entirely.
Daniel Votruba ·
In practice
When you ground a model, require the answer to cite its source. If it cannot name one or finds nothing relevant, it should say it does not know.
Read more: How to get into ChatGPT's answers: a practical GEO guide
Related terms
- Hallucination: A hallucination is a case where a model states plausible-sounding but false information.
- RAG (retrieval-augmented generation): RAG is a method in which a model first retrieves relevant documents, for example company knowledge, and answers from them.
- Large language model (LLM): A large language model is a model trained on a huge amount of text that predicts the next words and can write, summarise and answer.
- Semantic search: Semantic search finds results by the meaning of a query instead of matching individual words.
More terms in the group AI and automation
Guardrails · Generative AI · Function calling (tool use) · Human in the loop · Foundation model · Inference and latency · Fine-tuning · Jailbreak (AI) · Few-shot and zero-shot prompting · Explainable AI (XAI) · Machine learning (ML) · EU AI Act · MCP (Model Context Protocol) · Embedding · Model bias (AI bias) · Diffusion models · Model evaluation (evals) · Deep learning · Multi-agent system · Deepfake and synthetic media · Multimodal model · Custom GPTs and projects · Neural network · Copilot (AI assistant)
Terms in practice
Need help putting them to work?
Terms are one thing, putting them to work is another. We go through which of them pay off for you and say so plainly. You get a first opinion for free and we reply within 24 hours.
Book a call