Here is my list; this is based on my experience working with CouchDB, HBase and others
What is the tools/infrastructure enterprise use to scale their RDBMS?
What kind of strategies used to avoid (SPOF) due to DB failure?
What are the options available for DB2? SQL server? or fill your DB of choice
What is the possible trade off?
What happen to ACID properties to this situation?
What are the issues with using better commodity hardware?
This is very important as quest for solution should be based on existing problems and you need to understand that better to explore which would fit snugly.
What is their point of view on CAP?
What kind of applications can be easily developed, hard to develop using NoSql?
Is MapReduce recommended over SQL? What would be the productivity loss/gain?
How will it fit into virtualized environment? What is the impact of virtualization in terms of performance?
What are the potential issues when system goes away with mathematical relation model to document/Key value oriented model?
What is the theoretical limit on data size, data growth better served by NoSql?
Can NoSql be used to make backend/data services by pool of the hardware (commodity) by exposing select/update/delete/archive? And hiding RDBMS aspect from developers?
What is the roll of DBA in the NoSql world?
How is the transactional model like 2P commit/distributed commit fit into NoSql?
What is the impact of tools/framework/reporting framework developed specifically RDBMS in mind?
Enterprise security or product security impact using NoSql?
What would be the threat for NoSql from patent bullies like Oracle/Microsoft/others?
What is the perspective of vendors like SAP/IBM/Oracle and what is their take/offering on NoSql?
This questions will make you understand why nosql.