Download PDF

Scale AI > Case Studies > Goodcall Enhances Chatbot Performance with Scale Rapid's Text Annotation

Goodcall Enhances Chatbot Performance with Scale Rapid's Text Annotation

Technology Category

Analytics & Modeling - Natural Language Processing (NLP)
Analytics & Modeling - Robotic Process Automation (RPA)

Applicable Industries

Telecommunications

Use Cases

Chatbots
Speech Recognition

Services

System Integration
Training

The Challenge

Goodcall, a company providing businesses with intelligent phone agents, faced a significant challenge in managing and annotating the high volume of data generated by their chatbots. The chatbots, which use automatic speech recognition (ASR) to convert speech-to-text and AI analysis to interpret customer requests, required regular fine-tuning with real-world production data. However, the process of labeling this massive amount of data with high-quality annotations was time-consuming and resource-intensive. Furthermore, Goodcall was unable to match the scale of available data due to their in-house data annotation process. This meant that every piece of unlabeled data was a missed opportunity to improve their models. To enhance model performance and customer experience, Goodcall needed a scalable, sustainable approach for labeling large quantities of data.

About The Customer

Goodcall is a company that provides businesses with intelligent phone agents. These agents are essentially chatbots that customers can call to request information about the business’s hours, location, and products or services. The chatbots use automatic speech recognition (ASR) to convert speech-to-text and then employ AI analysis to interpret the person’s request and provide the appropriate response. The intent classification models that power Goodcall’s chatbots need to be regularly fine-tuned with real-world production data. Goodcall's goal is to improve the customer experience for their clients by enhancing the performance of their chatbots.

The Solution

Goodcall partnered with Scale Rapid to acquire high-quality text annotations. Scale Rapid's user-friendly tools allowed Goodcall to send large collections of samples for review and receive text classification labels in return. These labels, which describe the meaning of certain phrases, provided the data that software engineers needed to evaluate and improve their models. Goodcall also used these human-reviewed labels to check against the unlabeled items in their system, thereby expanding the variety of data they could use to improve their models. This process helped Goodcall improve the accuracy of a model’s text-to-speech transcripts and enhance a bot’s ability to correctly understand intent.

Operational Impact

Working with Scale Rapid has not only improved Goodcall's model performance but also enhanced the overall customer experience. By outsourcing the labeling process, Goodcall has been able to free up engineering time to tackle other problems. Additionally, using Scale Rapid’s text classifications, Goodcall has identified opportunities for improvement in their speech recognition system, uncovering limitations in their transcript and intent classification models. These text labels provide Goodcall with valuable insight into their chatbots, allowing them to quantify their models, identify issues, and assess why their models perform the way they do. This has enabled them to make intelligent decisions to improve their product and, by extension, the customer experience.

Quantitative Benefit