Generally speaking, you assign one developer to each adapter. This was, each developer can work on their own independent projects without worrying about any conflicts with another developer. Another developer (or one of the ones working on an adapter) would then create a "controller" project that would reference the adapter projects and call the logic it needs in those. This keep everyone separate.
By having a single developer on an adapter, that developer will become familiar with that application and know how to do all of the ancillary things that you are likely not automating as well as the nuances of the application (how it changes screens, accepts input, data format, etc...).