PRIZM AI

August 2026 · 4 min

How ASK PRIZM works

ASK PRIZM answers from knowledge a person in the organization has approved. It does not invent policy. Unknown questions are recorded, not guessed.

ASK PRIZM is the assistant on a PRIZM identity and on the public site. After a tap, a visitor can ask about the person or the business. On the website, a visitor can ask about the platform. The important part is not the chat window. It is the rule underneath it.

Answers come from approved knowledge. A draft, a transcript, or a repeated question can produce a suggestion. That suggestion is not live. Someone in the organization approves, edits, or rejects it. Only then may PRIZM speak it. The sequence is transcript → AI suggestion → human approval → published knowledge.

If ASK PRIZM does not know, it should say so and record the gap. That is how Chief and Command see where the business is being asked something it has not yet decided to answer. Guessing in the company's name is the failure mode this rule exists to prevent.

ASK PRIZM is not a general-purpose chatbot and not an unbounded model with the company logo on it. It is a governed path into the same knowledge Voice uses on the phone and Connect uses after a tap. One layer. One approval process.

On this marketing site, ASK PRIZM will answer from approved knowledge once the Origin API is connected. Until then, it says so. That is deliberate. PRIZM should not improvise just because a form is empty.