The recent release of ChatGPT and the announcement of Google Bard are generating a lot of buzz, a bit of controversy, and sparking conversation across the web. It appears that we are witnessing a new technology revolution, and just about everyone with an internet connection is excited to try these tools out. Here at Spitch, we are no exception. We have been eager to evaluate ChatGPT’s capabilities and to understand how those might benefit our own CX products.
Spitch’s AI-driven solutions are designed to help contact center providers to better understand their customers, and continuously improve the quality and cost-effectiveness of their services. The Spitch family of products and services currently includes: omnichannel (voice and text) conversational AI platform for customer and employee service support with seamless telephony integration, Virtual Assistants for dialog management, Voice Biometrics, Speech Analytics, a universal natural language understanding (NLU) engine, conversation transcription, and a constellation of other components and conversational AI technologies. ChatGPT on the other hand is setting new standards in NLU and content generation, and already transforming the way that people expect to interact with their digital devices.
We see this new technology as a potential source of new opportunities, both for us, our customers, and the industry in general. Whereas Spitch services focus on directing, collating, and understanding the customer journey, the ChatGPT solution provides an opportunity to augment those journeys with a new standard in information retrieval and natural interaction. Where ChatGPT provides general feedback and understanding, Spitch remains focused on custom-tailoring our solutions to specific customer and contact center needs.
Spitch conversational AI solutions provide a clear organization of AI-driven omnichannel customer content, but we can already see places where ChatGPT-like technologies may provide opportunities to augment or supplement our offerings:
Today we are just beginning to scratch the surface of what this new technology paradigm can do. However, we can also see some clear places where it is still lacking, and some where it might never fit the bill.
In particular, ChatGPT is a conversational engine, but it is not equipped or designed to provide clear and reliable references. Moreover, it is not designed to provide any kind of active linking or retrieval. It engages, it responds, and it draws on the vast reserves of a large chunk of the written knowledge of humanity, but it has no mechanism to link that knowledge or interactions to specific requests, specific tickets, or any specific customer data. It can hardly be used to reference individual customer conversations, or link records in a database, or correctly reference specific instances of prior information.
It is not (currently) tunable. Spitch solutions focus on fine-tuning deployments for specific customer needs. ChatGPT currently provides no mechanism to adapt its responses or restrict them to any particular or more limited domain.
It is not constrained by correctness. Spitch solutions focus on limiting responses to a customer defined domain of correct content. ChatGPT currently has no such content constraints and thus may sometimes respond with irrelevant or entirely incorrect content.
At Spitch we are always excited about the future of AI, and today we are particularly keen
to see how ChatGPT and others are going to influence that future going forward. We see a lot of synergies, a lot of opportunities, and a lot of room for us to grow into that future.
Spitch will certainly continue to bring the latest conversational AI technologies to our customers.
