Betting on LLM’s Intelligence: How Contextual Understanding Drives Outperformance

 LLMs are everywhere now. But in equity investing, adoption is still at an early stage as models are costly to run and hard to fine-tune for financial language. That’s where ProntoNLP Transcript Analytics comes in. Purpose-built for finance, it captures sentiment, context, and importance levels from earnings calls to generate differentiated stock selection signals.

This webinar provides a detailed blueprint on how investors can systematically leverage LLM-based transcript features to identify future outperformers.

Register Now 
If you can't make the live session, please register anyway to get access to the on-demand version.

Or click here to learn more about ProntoNLP. 

Key Webinar Takeaways:

• Contextual Understanding Drives Outperformance: A long-short strategy utilizing LLM-based sentiment analysis has outperformed a lexicon-based benchmark by 4.3% per year from 2010 to 2024.
  
  
• Financial Performance Is (Still) King: Among all events identified by LLMs, sentiment signals associated with financial performance actuals yielded the strongest long-short performance.
  
  
• Importance Level Matters: LLM-based sentiment signals regarding financial performance that are deemed highly important have outperformed those deemed unimportant by 6.9% per year on a long-short return basis.
• Recent Performance Remains Robust: Sentiment signals linked to financial performance and macroeconomic indicators have generated robust returns in recent years.