A question from a Twitter following:
Is it not overkill for an organization with fewer than 100 mailboxes and a single site to have a DAG implemented?
Database availability groups are the high availability building block for Exchange Server. The question of whether or not to implement a DAG should always be one that is approached in terms of business requirements instead of user count. A highly available email service may be very important to those 100 users, or even to 50 or 10 users. If the business need outweighs the cost, then why not deploy a solution that meets those needs?
Perhaps the organization already has a robust, highly available virtualization platform deployed, with reliable internet connectivity and a software licensing agreement that makes it trivial to deploy an additional Exchange VM to form a DAG.
Of course, the existence of Office 365 means that deploying a highly available on-premises may not be the most cost-effective approach. There's no correct answer because each customer and their associated costs and other factors are different, but you'll often find there's a sweet spot where it's most cost-effective to use Office 365 than to try and provide high availability on-premises.
But ultimately, the answer is no. It's not overkill if the customer requires that level of availability.