To accelerate the acceptance of new concepts developed under NextGen, the Shadow Mode Assessment using Realistic Technologies for the National Airspace System (SMART-NAS) testbed, which enables integrated examinations of NextGen or beyond-NextGen concepts under distributed environment, becomes critical to the Air Traffic Management (ATM) community. To support human-in-the-loop (HITL) testing for NAS-wide simulation using SMART-NAS testbed, this proposal addresses the feasibility of constructing an ATM-centric speech-enabled agent as a plug-and-play service of the SMART-NAS testbed. This service addresses the gap of HITL testing that is currently limited to small regions of airspace and few airports with a small number of controllers and pseudo-pilots. Leveraged from our prior development on noise-robust speech recognition system for the Navy and virtual agents for NASA to support HITL simulations, an infrastructure of ATM-centric speech-enabled agent will be developed. A feasibility demonstration of the speech agent as a service component of the SMART-NAS testbed will be provided by the end of the Phase I research. Phase II work will utilize the infrastructure built in Phase I to expand the speech-enabled agent to a full-scale prototype that supports HITL testing for NAS-wide simulation using the SMART-NAS testbed.