There is nothing about ADSL that requires the use of a VPN, so you could disable the VPN and rely on some sort of access control, such as a password, to provide rudimentary security. Whether this is a good idea depends on the nature of the traffic between the offices and what security experts call the threat model, the nature of the threats that you are trying to protect against.
A better solution may be to diagnose the cause of the VPN's failure. You don't say what kind of VPN the customer is using, so it's difficult to offer any specific advice, but it is often useful to take a packet capture of the VPN's establishment protocol (using etherpeek, tcpdump, wireshark or a similar tool), and analyze the resulting output. If your sites have dynamic IP addresses, it might be useful to ask your ISP if they have changed their DHCP lease policies lately, as this could be causing a change in one side's IP address that confuses its peer.
Learn how to connect sites with dynamic IP addresses across the Internet.
Dig Deeper on MSPs and cybersecurity
Related Q&A from Jon Snader
Learn how to set an IP address on the network interface of a FTP/Web/mail server when a client has only one public IP address. Continue Reading
To connect to a WAN remotely, your client can use a VPN client or a leased line. Learn the pros and cons of WAN connectivity with each option, such ... Continue Reading
When using ISA 2000, some users can lose connection to the network and experience packet loss even if the VPN client is still connected. Learn how ... Continue Reading