The sales tech is well primed to make the best usage of the current advances in Large Language models. While the pandemic accelerated the adoption of Software as a Service (SaaS) tools capable of recording, transcribing, and providing offline feedback on sales conversations, the LLM revolution that followed continues to be impactful for unlocking the insights hidden within those conversations at scale. In this talk we go over the learnings from our journey of leveraging text embeddings from state-of-the-art LLMs for sales use-cases involving retrieval and pattern mining. We show the performance of basic text embeddings and research methods for their application across various sales scenarios. Additionally, we delve into the advantages of tailoring embeddings to circumstantial usage cases and datasets, contrasting results against those of standard embeddings.
GeeCON 2024: Srikar Yekollu - Harnessing the power of LLM text embeddings
The sales tech is well primed to make the best usage of the current advances in Large Language models. While the pandemic accelerated the adoption of Software as a Service (SaaS) tools capable of recording, transcribing, and providing offline feedback on sales conversations, the LLM revolution that followed continues to be impactful for unlocking the insights hidden within those conversations at scale. In this talk we go over the learnings from our journey of leveraging text embeddings from state-of-the-art LLMs for sales use-cases involving retrieval and pattern mining. We show the performance of basic text embeddings and research methods for their application across various sales scenarios. Additionally, we delve into the advantages of tailoring embeddings to circumstantial usage cases and datasets, contrasting results against those of standard embeddings.
