Conversational Assistant Eva
Eva is an automated virtual assistant designed to engage in real-life conversations. It showcases the capabilities of artificial intelligence, specifically Natural Language processing (NLP). Eva was first used at the largest European technology event GITEX Europe in 2025 to answer questions related to Neurotechnology, and can be adapted for various applications, including exhibitions, company representation, and integration into the image of your startup or business.
The conversational assistant Eva can:
- Focus on specific topics. Eva can be trained to become an expert in particular subject areas, ensuring she provides accurate and relevant information on demand.
- Speak multiple languages. Currently, Eva can understand and respond in both English and Lithuanian, making her accessible to a wider audience.
- Follow distinct instruction sets. Eva's behavior and responses can be tailored through specific instructions, allowing her to adapt to different roles and interaction styles and give predetermined answers.
- Act as a virtual welcoming hostess for businesses. With her ability to engage in conversation and provide detailed information, Eva can serve as a digital "receptionist" for greeting visitors, answering common questions and directing people in a business environment.
How the conversational assistant Eva was built:
- Speech-to-Text (STT): This component transcribes human speech into text, allowing Eva to understand spoken queries.
- Large Language Model (LLM): The transcribed text from the STT is fed into the LLM, which then processes the input and generates a response. It serves as Eva's thinking apparatus, with instructions on how to answer questions.
- Text-to-Speech (TTS): To enable Eva to "speak", the LLM's text response is converted into an audio format.
- Lip syncing: Eva's avatar's lip movements are synchronized with the audio generated by the TTS, creating a realistic visual of her speaking.
- Directional sound: To make Eva's voice appear as if it's coming directly from her avatar, we used the Focusonics Model B speaker, which can only reach the listener in a localized zone, making it an immersive experience.
Get your custom conversational assistant.
Learn more about Eva, our interactive virtual assistants and other NLP technologies.