I would imagine the ability to act as a node in a fully networked and integrated ASuW/ASW force needs to be an absolute priority, so high speed and very secure data comms would be high on the list. If you see it, they need to see it, especially for things like land attack and over-the-horizon work.
Eliminating hosted aircrew would be my secondary consideration (boy I'm on thin ice now) - you have to feed them, provide ejection seats and scrabble boards, and they take up valuable fuel and equipment space. The technology isn't quite there yet particularly for rescue missions, but we are getting close. Autonomous functioning would be a must, because remote piloting links can always be interrupted.
It all depends if you're looking to design a lower risk platform using current technology, or if this is more of a forward look with the associated risks and costs of technology development. That needs to be one of the first questions defined by your project goals and deliverables.