Why Our AI Assistant Never Makes Things Up (and What Happens When It Doesn't Know)

One of the most common fears business owners raise when considering phone automation is simple: "what if the bot tells a customer something that isn't true?" It's a fair concern — many general-purpose AI chatbots, built without this specific risk in mind, will confidently state a service price, an opening hour, or a return policy that nobody ever actually gave them. It sounds fluent and convincing, but it's invented. For a real business's phone line, that's not a curiosity — it's a real risk: an upset customer, a wrong answer that gets repeated further, or a misunderstanding that ends up on your desk anyway.

The zero-guessing rule

Our AI Assistant works differently by design, not by accident. Every answer is based strictly on what actually exists in your business's knowledge base — information you entered yourself, or that the assistant gathered automatically by analysing your website during onboarding. If a caller asks something that isn't in that knowledge base, the assistant doesn't try to improvise. Instead, it honestly says it doesn't have that information right now, and that the business will follow up or add it.

What happens to that question

That question doesn't just disappear. Every unrecognised query is automatically logged in the client panel as a "knowledge suggestion" — with the exact wording of the question and a counter for how many times the same topic has come up. If five different callers ask the same thing (say, "do you allow pets"), you'll see one entry with a counter reading "5", not five separate notifications. One click — "Add to assistant's knowledge" — and that answer becomes a permanent part of the knowledge base, never a gap again.

Why this matters for trust

A customer who hears wrong information over the phone rarely calls back to correct it — they simply remember that the business gave them the wrong price or the wrong hour, and their trust drops. An assistant that says "I don't have that right now, but we'll check" instead of guessing sounds less impressive in the moment, but builds something far more valuable over time: credibility. It's the exact same instinct a good, honest employee would have — better to say "let me check and get back to you" than to guess.

A knowledge base that grows itself

In practice, this means your assistant's knowledge base gets more complete over time — not because you had to predict every possible question in advance, but because the system shows you exactly what's actually missing, based on real conversations. You don't have to guess what's worth adding — you see it in black and white in the panel, sorted by how often each question actually comes up.

If you want your business phone line to always tell the truth — even when that truth is "I don't know yet, but I'll find out" — that's exactly the mechanism that guarantees it.