I have personally serious issues with relying on a potentially buggy piece of technology to ensure safety on a group ride. Ensuring everyone in the group is safe is a fundamental human requirement of any group ride, regardless of whether it is a trail ride or on the road.
If you are familiar with ensuring safety in the workplace there is what is termed a hierarchy of controls, where there is a cascade of policies in place to ensure safety within a workplace (figure below).
The idea behind this hierarchy is that the control methods at the top of graphic are potentially more effective and protective than those at the bottom. Following this hierarchy normally leads to the implementation of inherently safer systems, where the risk of illness or injury has been substantially reduced.
Here, I would classify the smart phone application as PPE (personal protective equipment), the least safe option, it would only be trusted as a last ditch effort to keep someone safe after all of the other control elements fail.
We could eliminate the issue of lost riders by keeping the group size very small so that everyone is within sightline of one another.
If large groups cannot be avoided, we can substitute the risk by riding on low risk areas that everyone is familiar with if we have to maintain large groups. (Note this is slightly less safe from the first one as someone could still get injured and abandoned.)
We can have engineered controls, by breaking a large group into subgroups with a leader and sweep within each smaller group. (Note this is less safe again, as a whole subgroup could be lost.)
We can having administrative controls by having leaders of the subgroup would then report to the primary ride leader at checkpoints. (If we determine a group is missing, we are already behind the eight ball, hence why it is lower on the list.)
The application would come last in this list. The application would have to be functional (cellular and GPS functioning), the rider would have to be able to use it (i.e, no injury, not panicked, sufficient charge in the battery), and the rider would have to be capable of self-extraction (i.e., reading maps, picking a safe route across the terrain).
This element of risk may be why no one has created such an application as it could be a serious liability risk for any company that does so. To ensure safety you have to ensure that smartphones have cellular access and that GPS signal can be always be maintained, that the riders can use the app and can self-extract successfully. That is a long list of conditions. I don't know where you mountain bike, but in British Columbia Canada few if any rides have cellular and GPS signal making this more of an interesting concept rather than a viable safety alternative.