Lauren Kunze, CEO of Pandorabots, discusses the growth of Pandorabots as a leading chatbot development platform with collaborations with global brands. The podcast explores the versatility of chatbots, their role in combating loneliness, and the challenges of building good chatbots. It also covers integrating chatbots in companies, breakthroughs in natural language processing, and the rise of virtual chatbots and virtual talents.
Building chatbots remains a complex task, with a focus on training bots for specific domains to reduce human workload.
Testing, quality assurance, and human oversight are crucial for improving bot performance and preventing biased responses.
Deep dives
Pandora Bots: A Platform for Building Chatbots
Pandora Bots is a chatbot development and hosting platform with a large user base of registered developers. It has built over 350,000 chatbots for various applications, including work with global brands and Fortune 500 companies. The platform has processed over 80 billion conversations between bots and humans. Bots have become more popular due to automation needs during the COVID-19 pandemic, including increased customer service traffic and addressing loneliness. The use cases for consumers have expanded, with a focus on social companionship and wellness.
Building Good Bots is Challenging
While building chatbots has become more popular, it remains a complex task. Natural language understanding, a field in artificial intelligence, is still a challenge. It is difficult to build a good bot that fully understands and responds to human conversation. Many companies focus on building bots for specific domains or tasks, such as customer service. By training a bot to handle a limited set of questions, businesses can reduce the workload on human agents. However, ongoing training and tuning are necessary to improve the bot's performance.
The Importance of Testing and Quality Assurance
Testing and quality assurance are crucial in bot development. Companies employ various methods, including unit and regression tests, mock conversational data, and log review and analysis. User testing and observation play a significant role, helping to identify conversation flow issues, drop-off points, and refining the user experience. Human oversight is necessary to prevent bots from learning inappropriate or biased responses based on user interactions. Regular monitoring and updates based on real-time data are required to continuously improve bot performance.
The Future of Bots and Virtual Beings
The future of bots is likely to involve avatars, combining speech, text, computer vision, natural language understanding, and emotional expressions. Advances in technology, such as deep fakes and generative adversarial networks, contribute to creating lifelike faces for bots. The use of virtual beings, like virtual influencers and digital humans, is becoming more common, especially in Asian markets. Bots and virtual beings offer new possibilities for entertainment, customer service, and live events. However, challenges related to training data, bias, security, and achieving natural human-like interaction must be addressed.
In this episode of the Data Exchange I speak with Lauren Kunze, CEO of Pandorabots, a widely used platform for building chatbots. About four years ago I attended Bot Day in San Francisco, and at the time, chatbots were very much in the news. Today, chatbots are used across many industries and use cases, and on many types of devices. Lauren Kunze and Pandorabots have been at the forefront of many important developments in the conversational applications space. They assist many enterprises build and deploy bots, and they also create leading edge chatbots like Mitsuku.