Conversational Voice Assistant for Inventory Management

Inspiration

Make data and processes accessible and simple for every Enterprise user without learning complex app, dashboards UI and totally hands-free. Simply ask and get things done.

Today, enterprises have multiple siloed applications with their own interfaces that requires time and effort to learn. With our approach Voice AI remains a single interface to these various siloed backends thus giving the user access to information at rapid speed without any learning. They can converse in their regular language, fetch data immediately and take decisions faster.

What it does

Using our patented Speak to IoT platform, we have built a Conversational Voice Assistant that can communicate with backend Database providers that handles Warehouse Inventory.

How we built it

Custom voice model for English US locale.
Contextual awareness to handle various user intents.
API integration with SingleStore DB
GitHub README captures additional details.

Challenges we ran into

Extracting data from multiple siloed applications to provide a holistic end-user experience.

Accomplishments that we're proud of

Using our platform we were able to achieve the entire end-to-end solution in less than a week.

What we learned

Handling of complex user queries and navigation using Voice is extremely hard especially when dealing with multiple siloed backends. Accuracy and precision are very crucial to ensure end-user experience is not suffered. Using our current approach we are able to achieve 90% accuracy and to achieve more we need access to larger data-sets. Over time the accuracy will improve further based on closed loop automation.