this article presents the experimentation and results of applying the Adaptive Automata technique to a Spanish voice synthesizer looking for its improvement. It shows a review of the design and implementation of the proposed solution, both aspects are fully explained in a previous work. It also highlights the formal methods available to measure the improvement on the voice quality, specially the naturalness feature. Lastly, it presents the results of experimentation with users for both versions, the original and modified system. These results show that significant improvement can be made in understanding the meaning of read texts by incorporating linguistic concepts through the use of adaptive technology. © 2010 IEEE.