/dqc/media/agency_attachments/3bO5lX4bneNNijz3HbB7.jpg)
Follow Us/dqc/media/post_banners/wp-content/uploads/2023/03/Dattaraj-Rao.jpg)
Dattaraj Rao, Chief Data Scientist at Persistent Systems talks about conversational AI in this interaction.
What are the key trends and developments in the field of conversational AI?
Dattaraj Rao - Conversational AI is being revolutionised by large language models (LLMs) and generative AI. While traditional chatbots were good at finding things and showing them almost as-is to the end user, we can achieve a human-like paraphrasing and conversation capability with LLMs. LLMs are very good at taking a response and rewording it in a human-like manner. Along with LLMs, better interfaces and omnichannel integration with applications like Microsoft Teams, Salesforce, Facebook, WhatsApp, Alexa and more will make conversational AI accessible to all. We will see easy context-switching to continue conversations from one channel to another. Instead of each application giving us a custom conversation interface, we will see more generalization and continuity of conversations.
What is the future outlook for Conversational AI in and beyond 2023?
Dattaraj Rao - We will see smarter virtual assistants that understand the domain context and omnichannel interfaces like chatbot, voice and even virtual reality, with metaverse. Besides giving human-like responses, the assistants will provide the ability to understand the user’s intent and connect to back-end systems like ERP and CRM to get relevant information. In addition, omnichannel interfaces will provide convenience and continuity of conversations across platforms.
How have you grown in conversational AI in the past one year?
Dattaraj Rao - Persistent has worked with major BFSI, Healthcare and Industrial clients to build next-generation conversational interfaces. Our clients have increased and we have increased customers in the BFSI domain, particularly in Europe.
Do you integrate ChatGPT and if yes, in what ways?
Dattaraj Rao - Our AI Research Lab actively investigates large language models like GPT3.5 and ChatGPT; we are building applications for Customers using this technology. ChatGPT helps us build better conversational interfaces and provide human-like responses. We are helping customers integrate domain knowledge bases in banking and insurance with the ChatGPT interface to provide context and meaningful information to end users. For example, in insurance domain, we understand existing coverage for customers and use an LLM like ChatGPT to recommend better coverage plans and uncovered risks specific to that customer. All this is done in a standardized manner using ontologies that define the structure of the data.