Internet connectivity with many single points of failure is adequate for the majority. You wouldn't be reading this far if it were adequate for you.
Consider this analogy: many companies delegate the management of their Internet connectivity to their ISP. In the same way, many delegate the generation and delivery of electricity to a single local power utility.
But some firms have their own generators. Some have connections to multiple utilities.  Using two ISPs for reliable Internet connectivity is very much like using two power sources to provide relaible AC power. (The fancy equipment needed to switch a load between two power sources is analogus to routers running BGP that transfer Internet traffic from one ISP to another.)
The details of your need for reliable Internet connectivity are probably quite different from anybody else's. Hence, I believe that you'll best be served by investing the time to learn how to build the best solution that meets your needs instead of leaving these details to a delegate.
Identifying and minimizing single points of failure in your current and proposed network architectures
Understanding the costs and benefits of alternative architectures
Choosing the architecture for your network
Choosing routers and Internet Service Providers appropriate for your architecture
Coordination with the parties involved
Establishing policies for your network
Configuration of your equipment
I once toured an AM radio station's transmitter facility. Its 50,000 watt signal covered over a dozen states with a potential audience in the 10's of millions. Hence, reliably transmitting their signal was important.
They could draw power from three different distribution grids. Think about it. What is the output of a radio station when its transmitter has no power? Silence.
Copyright © 1999-2000 by Robert A. Van Valzah