Glossary
AI that
Knows what it doesn't know.
An AI that recognizes when it doesn't have a confident answer, says so plainly, and points to what's missing. The opposite of confident hallucination. Honest about its limits, useful because of it.
Most AI tells you something. Confidently. Even when it shouldn't. The smartest answer is sometimes 'I don't know yet.'
What this looks like in practice
Confident answers, wrong answers.
A generic AI assistant answers every question with the same confident tone, whether it's quoting a verified source or hallucinating something plausible. By the time someone catches the wrong answer, the decision is already made. Reliable AI says 'I don't have that' instead of inventing one.
Gaps you didn't know you had.
A knowledge base only tells you what's in it. It can't tell you what should be there but isn't. Honest AI notices when the questions keep coming and the answers keep being absent. The gap surfaces as a thing to fix, not a silent failure mode.
Trust that grows instead of erodes.
Every confidently wrong answer is a trust withdrawal. Every honest 'not sure' is a trust deposit. Over a few months, an AI that admits its limits becomes more useful than one that fakes confidence, because people actually act on its answers.
Honesty is a feature, not a flaw.
Lumen tells you when an answer is grounded in your documents, when it's a best guess, and when the knowledge is missing entirely. The result is an AI your team can act on without second-guessing.
Bring your knowledge.Simply ask.
CASA Certified · SOC 2 Infrastructure · GDPR-Aligned · Enterprise Ready