Speech intelligibility is a key factor in conference room acoustics, affecting the listener’s experience and communication effectiveness. Traditional acoustic design relies on objective parameters, often neglecting perceptual validation of both acoustic and aesthetic features. This study explores how virtual reality technologies can support the acoustic design process. Two acoustic treatment solutions for the conference room of the Museo Egizio di Torino were compared: a (1) seamless plaster-based system and (2) parametric felt baffles. Geometrical Acoustic software Ramsete was used to predict their effectiveness, and to generate third-order Ambisonics auralisations. A photogrammetric 3D model was used for visual renderings. Both were used for speech intelligibility tests conducted in the Audio Space Lab of Politecnico di Torino, a small listening environment equipped with a High-Order Ambisonics sound reproduction system and an Oculus Meta Quest 2. Listeners were also asked to express their preference on the visual scenarios obtained with the two treatment solutions. The results showed a preference for the second solution, despite no significant acoustic differences.