Predictive voice assistant

The other day, I came across an intriguing article on Medium from Sensory about voice assistants and the different methods to initiate conversations with them. The article highlighted three primary approaches:

Traditional Wake Words: This approach uses locally interpreted wake words like "Alexa," "Google," or "Siri."

Push-to-Talk: Devices like the Humane pin and Rabbit utilize a physical button to start interactions.

LLM Smart Assistants of Today: Examples like ChatGPT-4o can sustain a conversation once initiated.

Given our experience over the past few years developing a voice assistant for virtual travel guidance (Guide My Trip), we've explored and experimented with each of these methods. Initially, we experimented with wake words, desiring a locally interpreted solution. However, the high cost of effective products presented a significant barrier for our startup. We hope that tech giants like Apple and Google will eventually provide their wake word services to app developers for free.

As a practical alternative, we implemented a push-to-talk system, allowing users to press a key to start a conversation. Although this method is not ideal—especially for users who might be driving—it was the most viable option at the time. Once the conversation starts, the goal of our app functions similarly to ChatGPT-4o, albeit without a streaming interface (for now).

Furthermore, we innovated a predictive approach to enhance user interactions. Guide My Trip leverages the mobile device's location, direction, and speed to anticipate the user's destination in real-time. By informing users about interesting places ahead based on their preferences, we can initiate relevant conversations. Users have previously indicated their interests, enabling us to provide timely and pertinent information about nearby attractions.

This predictive capability allows us to initiate and guide conversations about these places, offering users control over the dialogue flow. I firmly believe that predictive behaviour will be a crucial element in the future of voice assistants. These assistants will not only collect and analyse data about our behaviour but also proactively assist us throughout the day.

Incorporating predictive behaviour into voice assistants opens up numerous possibilities. For example, imagine a voice assistant that knows you have a meeting at 3 PM and proactively reminds you to leave early due to heavy traffic on your usual route. Or consider a virtual travel guide that not only suggests popular tourist spots but also uncovers hidden gems based on your past interests and travel patterns. The potential for creating deeply personalized and anticipatory experiences is vast.

Moreover, the ability of voice assistants to learn from user interactions and improve over time is another exciting frontier. As these systems become more adept at understanding natural language and contextual cues, their usefulness will extend beyond simple commands and queries. They will become true companions, seamlessly integrating into our daily lives to

In summary, the evolution of voice assistants—from traditional wake words to push-to-talk and advanced conversational AI—highlights a significant shift towards more intuitive and predictive user interactions. The future of voice technology is bright, with endless possibilities for making our lives easier and more connected. As we stand on the cusp of this exciting era, we are committed to pushing the boundaries of what voice assistants can do, ensuring that Guide My Trip remains at the forefront of innovation.

I would love to hear your thoughts and experiences with voice assistants. Have you tried any of these methods in your projects? How do you see the future of voice technology evolving? Feel free to share your insights and join the conversation!

Finally, we are excited to announce that Guide My Trip is currently in beta testing. We invite you to download the app from the App Store or Google Play Store and experience our innovative voice assistant firsthand. Your feedback will be invaluable in helping us improve and perfect the app. Thank you for being a part of this journey with us!