Questions people actually ask
The ones that matter.
What stops it saying the wrong thing to a customer?
You decide what it can send on its own. Your hours, your warranty terms, where you are today – no permission needed. Anything that quotes a price, promises a date or touches money waits for your yes. And when a question falls outside what it knows, it does not guess. It stops and hands you the whole thread. The worst case is that a customer waits for you, which is what happens today anyway.
Can I change it myself later, or do I have to call you?
You change it yourself. Write it the way you would explain it to a new hire – no weekend callouts under two hours any more – and it applies from the next message on. Prices, policies, hours, wording. Call me for something bigger, but you should never need me for a price change.
Does it run on your machines or on mine?
Either, and we make that choice together before anything is built. Frontier models are the most capable AI there is. They run on the provider's servers, you are billed for what you use, and they improve as the models do at no cost to you. Local means the same system on a Mac Studio I supply and set up in your office, so nothing leaves the building and there is no metered bill. It suits businesses under about five people. Past that, frontier is the better buy and I will say so.
What does local actually mean for my patient or client data?
It means the AI runs on a machine in your building and your records never leave it. That is the strongest privacy position available and the right one for dental, legal and accounting work. It is not by itself the same as being HIPAA compliant – there is more to that, and I will tell you exactly what is and is not covered before you spend anything.
What happens when a new channel comes along?
That is the part that keeps this from going stale. When your customers start reaching you somewhere new, the same system takes it on. It already knows your business, so it is a connection rather than a rebuild.
Is this going to replace my staff?
No – and I would tell you if it were. It answers the channels nobody can get to, which today means the ones that sit unanswered until you are back in the truck. If you have been putting off hiring someone for the phones, this is often what you do instead. For the people you already have, it takes the repetitive messages off them so they do more of what you hired them for.
What happens when it breaks and I have no idea how to fix it?
You text me. Nothing I build is allowed to be a single point of failure – every system has a manual fallback your staff already know, because it is how you work today. Then you and I get on a call and I fix it.
If we stop working together, what happens to everything it has learned?
It stays with you, in plain files you can open without me – the knowledge, the settings, the history. That is a term in the agreement, not a promise. I would rather you could leave easily and choose not to.
What actually happens after I say yes?
Four steps, all of it remote – I do not need to be in your building. We agree exactly what it will do and price it as one number before any work starts. I build it against your real work, your price list, your invoices, the email you send after every job, so you prepare nothing and there is no discovery call. It goes live one channel at a time, the safest first, alongside how you work now, and nothing is retired until you have watched it handle real work. Then it is yours, with recorded training, a one-page sheet per system, and the controls to change it yourself.
What if it does not work for my business?
You will know early, because you will have watched it run on your own data before it touches a single customer. If it is not right at that point, you pay for the time spent and we stop. That is a better outcome than launching something you do not trust.