Short answer: it says so, and takes a message. ZeroMiss answers strictly from a knowledge sheet your clinic approves — services, rates, hours, billing, policies. If a caller asks something outside that sheet, the AI doesn't improvise an answer. It tells the caller it'll have your team follow up, captures the question, and moves on.
Why This Is the Right Default, Not a Limitation
A generic chatbot that confidently invents a plausible-sounding answer is a liability at a health clinic. A caller asking "does my insurance cover this" or "is this treatment right for my condition" deserves an accurate answer or an honest "let me have someone get back to you" — not a fluent guess that happens to be wrong. ZeroMiss is built to prefer "I don't know, but I'll find out" over sounding confident.
What "Outside the Knowledge Sheet" Actually Looks Like
- Clinical questions: "Do I need an MRI for this?" — always redirected to your team, never answered by the AI.
- Coverage specifics you haven't documented: if a specific insurer or plan isn't in your approved sheet, the AI says so rather than assuming.
- Anything about a specific patient's history or file: the AI has no access to clinical records and doesn't pretend otherwise.
- Edge-case policy questions: unusual cancellation or rescheduling scenarios not covered in your standard policy get taken as a message.
What the Caller Actually Hears
Something like: "That's a great question — let me make sure you get the right answer. I'll pass this along to the team and someone will follow up." Not a hedge disguised as an answer — a clear handoff, captured as part of the call summary your staff sees.
How Your Team Sees These Follow-Ups
Every "I don't know, taking a message" moment is captured in the structured summary your team receives after the call — what was asked, in the caller's own words, alongside their contact details. Nothing gets silently dropped just because the AI couldn't answer it on the spot.
Frequently Asked Questions
Can we expand the knowledge sheet over time?
Yes — as you notice recurring questions the AI is taking messages for, add them to the approved sheet so future callers get an instant answer instead.
Does the AI ever try to reason its way to an answer using general knowledge?
No — it's constrained to the approved knowledge sheet for clinic-specific facts. It won't pull in outside information to answer a question about your rates, policies or services.
What about general health questions unrelated to booking?
Redirected to your team every time. The AI's role is reception and booking, not health information — even for questions that feel harmless.
How do we know how often this is happening?
Your call summaries show every instance, so you can see patterns — if the same question keeps coming up, that's a signal to add it to the approved sheet.
Bottom Line
An AI receptionist that never says "I don't know" isn't being helpful — it's being risky. ZeroMiss is built to take a message instead of guessing, which is exactly the behaviour a health clinic actually needs from something answering its phone.