In this podcast episode, Fergal Reid, Intercom's VP of AI, dives into the rapid evolution of artificial intelligence, particularly large language models (LLMs), and their game-changing role in customer service. He points out that while LLMs operate on a seemingly straightforward principle of token prediction, their capacity to grasp context and tackle intricate tasks is truly remarkable. Reid showcases Intercom's AI chatbot, Finn, as a prime example of effectively utilizing LLMs, emphasizing the creation of a strong, "industrial-strength" system that sidesteps the limitations of simpler, less effective models. He advocates for a data-driven, scientific approach to developing AI products, highlighting the importance of continuous improvement and excelling in a few key tasks rather than spreading efforts too thin. Looking ahead, Reid envisions a future where AI chatbots are essential, adeptly managing complex customer support challenges on their own and even working alongside human agents in a proactive manner.
Sign in to continue reading, translating and more.
Continue