handly.app ↗ | is an app solution that connects craftsmen with clients, was facing a challenge with managing a high volume of support requests from the craftsmen that were received via email. To address this challenge, we proposed a technically complex solution that involved embedding support resources and documentations into a vector database and using a state-of-the-art natural language processing model to provide quick and accurate support to the craftsmen.

To begin, we worked closely with Handly's support team to identify the most common issues faced by the craftsmen and developed a comprehensive knowledge base that contained all the necessary support resources and documentations. We then utilized advanced vector database technology to store and organize the information in a way that could be quickly and efficiently queried.

Next, we used a state-of-the-art natural language processing model, specifically the cutting-edge GPT-3 model, to understand the support requests received via email and provide relevant responses from the knowledge base. The GPT-3 model was fine-tuned to ensure that it provided accurate and helpful responses to the craftsmen.

Once the GPT-3 model was ready, we integrated it with Handly's support system. Whenever a support request was received from a craftsman, the GPT-3 model would analyze the request, search the knowledge base, and provide the most relevant response as the first layer of response. If the craftsman's issue was not resolved, the support team would step in and provide additional assistance.

The implementation of the solution led to a significant reduction in the overall manual support responses required from the support team. Within the first month of implementation, the system was able to provide accurate and helpful responses to 74% of the support requests received from craftsmen. This led to a 40% increase in the support team's productivity, allowing them to focus on more complex support requests that required their expertise.

In conclusion, our technically complex solution successfully addressed Handly's challenge of managing a high volume of support requests from the craftsmen that were received via email. The integration of the vector database and GPT-3 model provided a quick and accurate first response layer, allowing the support team to focus on more complex support requests. This project showcases the potential of advanced AI-automations to improve business processes and increase productivity.