Alex Kvamme, CEO of Pathlight, discusses the innovative use of large language models (LLMs) in customer service, the challenges of integrating multiple LLMs, tips for efficient integration of LMS into company operations, and the importance of diversifying language models in product development.
Pathlight utilizes large language models (LLMs) to perform deep human-level analysis of customer conversations and extract valuable insights for executives.
One challenge for Pathlight is fitting long conversations into the context window of LLMs, but they use techniques like compression and retrieval augmented gen to handle this.
Deep dives
Pathlight: Using LLMs for Customer Conversation Analysis
Pathlight, a generative AI native conversation intelligence platform, uses LLMs to analyze customer conversations and extract valuable insights for executives. Before LLMs, sentiment analysis and standard NLP techniques were limited and only provided a cursory analysis. With LLMs, Pathlight can perform deep human-level analysis of every conversation, identifying key information, answering questions like why customers cancel, and detecting emergent trends. While currently focused on text, Pathlight is experimenting with multimodal data and exploring potentials of audio analysis.
Overcoming Challenges in Context Window Limitations
One challenge Pathlight encountered was fitting long conversations into the context window of LLMs, which have token limits. They use various techniques like compression, Lang chain-driven map reduce, and retrieval augmented gen to handle conversations that exceed the context window. However, their primary focus is on throughput and processing millions of conversations in real-time. Scaling and managing the growing volume of conversations and API limits are other challenges they face.
The Future of LLMs in Customer Service
Pathlight believes LLMs will transform the customer service field and revolutionize how companies interact with customers. LLMs offer the potential to automate tasks, increase engagement through chatbots, and provide comprehensive insights to support faster decision-making. By harnessing LLMs for customer conversations, companies can scale faster, improve customer outcomes, and gain a deep understanding of customer needs, enabling transformative changes across various industries.
In this episode, we're excited to welcome Alex Kvamme, CEO of Pathlight, for an in-depth discussion about the innovative use of large language models (LLMs) in transforming customer service. Discover how Pathlight is pioneering new frontiers by integrating multiple LLMs to tackle rate limits, cost, and performance challenges. Alex will share insights into the creation of a unique LLMOps pipeline, a groundbreaking approach that has significant implications for companies like CLEAR, Wine Enthusiast, and Stitch Fix. This conversation promises to be a treasure trove of insights for anyone interested in AI's role in enhancing customer experiences and the technical intricacies of managing advanced AI systems in a business context. Join us as we explore the cutting-edge of customer engagement technology with one of its leading innovators.
Get the Snipd podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share & Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode