Skip to content

Prompting

Grounding prompt

Instructions that require an AI answer to stay tied to supplied sources or evidence.

Example

The prompt requires the answer to rely only on the attached policy and cite the relevant section.

Why it matters

It helps keep replies within an agreed evidence base rather than inviting unsupported additions.

You'll hear it

“Use only the attached policy, and say when it doesn't answer the question.”

Can you control it?

Yes

Yes. You can often use or adjust a grounding prompt directly when the product supports it. The available options and their effect still depend on the tool.

Common questions

Does the instruction make an unreliable source trustworthy?
No. An answer can follow a source faithfully even when that source is wrong.
Can the AI still add unsupported details?
Yes. The instruction reduces ambiguity but does not guarantee that every sentence follows it.
What if the supplied material is incomplete?
A well-handled answer should acknowledge the gap rather than quietly invent the missing information.

AI chatbot

Still have questions?

Ask the AI chatbot about this term or another AI concept.

Up to 500 characters.

Ask LATHIC about AI. Relevant glossary entries may be included.

Your question, the glossary entries it matches, and a rotating pseudonymous identifier go to Microsoft Azure’s OpenAI service through Vercel AI Gateway to generate an answer. Zero retention and no training are required of the provider, and LATHIC does not save your question or answer. Privacy Notice