Your data
Your business, and what we do to look after it.
You are about to let a computer near your books, your customers and your money. Here is exactly what that means, in plain terms, including the parts where we are not as grown up as a bigger company would be.
Where your information goes
Nowhere else.
Most worries about AI come down to one question: who else gets to read this? For us the answer is nobody, because we did not borrow somebody else's AI to build these products.
- The AI is ours. We do not send your information to another company to be processed, so there is no third party to have a policy about.
- It runs on our machines. Not rented space in somebody else’s cloud. We know where the computers are.
- Your data is never training material. Not by default, not quietly, and not as a line added to the terms one day.
- Your conversations are yours. What you asked the assistant, and what it did about it, is part of your record. You can export it and you can delete it.
- Nothing is shared between customers. The assistant working on your account can only ever see your account.
- Nothing changes under you. We do not swap the AI for a different one halfway through your month.
The practical side
Kept apart, locked up, and backed up.
The boring parts, which are the parts that matter when something goes wrong.
- Kept separate
- Your business is walled off from every other business in the database itself, not by a setting somebody could get wrong.
- Who at our end can look
- Named people, only when you have asked for help, and it is written down.
- Locked up
- Scrambled on the way to us and scrambled while it sits on the disk, including in the backups.
- Backed up
- Kept in a second place, and we practise putting it back. A backup nobody has ever restored is a hope, not a backup.
The brakes
What the assistant is simply not allowed to do.
You cannot stop somebody sending your business a cleverly worded email. What you can do is make sure it does not matter, and that is where we put the effort.
Every job the assistant can do is filed by how much harm it could cause, and that filing lives in our software rather than in the AI. So a message that talks the assistant into attempting something expensive still lands on your screen as a question, because the check was never part of the conversation.
If you have connected your own software to your account, it is held to this more tightly than we hold ourselves. Anything coming from outside that would change your data waits for a person, every time.
- Deleting anything with history attached to it
- Reopening a month you have already closed off
- Sending money, promising money or refunding money
- Sending anything to somebody outside your business
- Giving another program access to your account
This list is fixed. Loosening your other settings does not shorten it.
Being straight with you
Things a bigger company would have, and we do not.
We would rather you found this out here than three months in.
- We have not been through a formal security audit. If your company requires a SOC 2 report before it can sign anything, we cannot give you one today. Ask us where that work is and we will tell you honestly.
- We are not set up for health records or card numbers. These products are not built to hold them, so please do not put them in.
- Nobody outside has tried to break in on our behalf. We have not paid for an independent security test yet.
- We are small. There is no security team watching screens overnight. There are a few people who answer the address below quickly and will tell you the truth if something has gone wrong.
- No uptime guarantee on free plans. Paid plans promise what their own pricing page says and nothing more.
Found a problem?
Tell us and we will fix it.
We have no bug bounty to pay you from, and no lawyers to send at you either. If you have found something, we would genuinely rather know.
Questions people ask
Is my data used to teach your AI?
No. We train on our own material and on work we do ourselves. Using a customer's information would need a separate agreement that you signed and could cancel, and we have not asked anybody for one.
Does anything go to OpenAI, Google or Anthropic?
No. We built and trained our own AI and it runs on our own computers. Your books and your customer emails are not sent to another company to be read.
What happens to my data if I stop using you?
You take it with you. Export everything first, in files that other software can open, without asking us. Then we delete it when you tell us to. It clears out of our backups over the following weeks rather than instantly, because that is how backups work.
Could someone trick the AI into doing something bad?
They can try. Someone could email your support inbox with wording designed to confuse it. What they cannot do is get past the brakes: spending money, emailing customers and deleting records are held for a person no matter how the request was phrased. That check happens in our software, not in the AI, so there is nothing to talk round.
Can your staff see my books?
Only when you have asked us for help, and only for as long as that takes. It is logged and you can ask to see that log. We do not browse customer accounts out of curiosity.
Got a questionnaire from your accountant or your bank?
Send it over. We would rather fill it in properly than have you guess from a web page.