I hope you've thought about security. There should be a simple point-and-click way for agents to block call requests from a single end-user IP address or a ran

I hope you've thought about security. There should be a simple point-and-click way for agents to block call requests from a single end-user IP address or a range of IP addresses. Otherwise you'll get a DoS attack from the first bored teenager who decides that clicking "Call support" repeatedly on your client's website is more fun than making prank calls. You should probably use SSL for something like this -- at least the business end that actually hosts the scripts and handles the DOM exchange -- but I'm sure you've already thought of that.