Security and abuse protection

Built to stay calm under spam, bots and abuse

Any chat box on the internet attracts bots and people trying to misuse it. AskAI has protection built in: a quiet human check, an automatic shield that steps in when something looks wrong, and limits that keep costs predictable.

At a glance

  • Only works on websites you have registered
  • Quiet human check — most people never see it
  • Automatic shield that switches to a limited mode under attack
  • Spam protection on the contact form
  • Per-visitor and per-website limits keep costs predictable
  • Refuses attempts to make it ignore its instructions

Only on your website

Your assistant only answers on the websites registered to your account. Copy the script to another site and it simply does not appear there.

Real people, not bots

Before the assistant writes an AI answer, it runs a quick human check in the background. Most visitors pass without noticing. When the check needs a click, it appears inside the chat with one clear line, and the visitor’s question is sent as soon as they are done.

An automatic shield

The assistant watches for signs of abuse: floods of messages, repeated questions from many places at once, attempts to make it ignore its instructions, spam leads. A single misbehaving visitor is slowed down or paused without affecting anyone else. If a whole website is targeted, it switches to a limited mode for a while: visitors can still read answers from your own pages and reach you through the contact form, but no expensive AI answers are written. It returns to normal on its own, and you get a short email if it happens to your site.

Predictable costs

Limits per visitor and per website stop anyone from running up AI usage on your account. Simple greetings and contact requests are handled without using AI at all.

Fair and careful answers

The assistant answers from your website and your own written answers, not from guesses. It does not discriminate or make assumptions about people. For real estate, it follows fair-housing rules: it never describes areas by who lives there.

Questions

Will the human check annoy my visitors?

Most people never see it. When it does need a click, it is one tick inside the chat and their question is sent automatically after.

What happens if my site is attacked?

The shield switches your assistant to a limited mode for a while — answers from your own pages and the contact form still work — then returns to normal by itself. You get an email when it starts and when it ends.

Can someone use my assistant to run up costs?

Per-visitor and per-website limits are built in. In an unusually busy month the assistant keeps answering from your own written answers and website rules, and we email you — no surprise bills.

Put an assistant on the website you already have

One line of code. CA$10/month launch price. Office AI included. No card needed. Then CA$10/month launch price.