Availability Groups provide you the ability to scale out reads via secondary replicas (end user reads, like reporting workloads, as well as things like backup and DBCC). They do not offer anything new that lets you scale out writes.
Your options there are still things like merge and peer-to-peer replication, or manual sharding or distributed partitioning (where the app knows which database to write to and read from). I don't know to what extent (if at all) any of those solutions will support seamless failover. That might be better achieved with manual partitioning and Availability Groups or even mirroring - but again your app would have to understand that there is more than one database and they are distinct and separate. Availability Groups themselves don't provide anything to distribute the writes automatically.