Glossary
Source citation.
The practice of attaching a verifiable reference to every claim an AI answer makes, pointing back to the specific passage, row, or relationship in the underlying data the claim came from. Lets the user check the answer instead of trusting the model.
"Trust me, I'm an AI" stopped being acceptable around the time generative models started confidently making things up. A claim without a citation is a guess that sounds confident.
What this looks like in practice
What real source citation looks like.
Not "according to your documents." Not "based on company knowledge." The actual passage, with the document name, the section, and ideally the exact sentence the claim came from. Clickable, scrollable, verifiable. Without that, "the AI said so" is the only justification, and that's not enough.
Why "I synthesized this from multiple sources" hides bugs.
A common shortcut: the model paraphrases across several documents and gives one composite citation. Sounds thorough. Hides whether each part of the answer is actually grounded or partially invented. The fix is per-claim citations, not per-answer ones. Most products don't go that deep because per-claim verification is significantly harder than per-answer.
The user benefit.
When you can click a citation and read the actual source, you can decide for yourself whether the answer is right. The AI stops being a black box and becomes a research assistant. Trust shifts from "I trust the model" to "I trust what I just verified for myself."
The footnote is the trust signal.
Simply Asking cites the specific passages every answer pulled from. You don't have to trust Lumen. You can verify the answer in one click. That's the difference between AI as oracle and AI as research assistant.
Bring your knowledge.Simply ask.
CASA Certified · SOC 2 Infrastructure · GDPR-Aligned · Enterprise Ready