Bar Segal

Title: Bar Segal and Guided Dynamic Construction of LLM Prompts

Introduction

Bar Segal is an inventor associated with Tel Aviv, Israel, and Microsoft Technology Licensing, LLC. Her patent work addresses how large language models can produce more relevant business correspondence while reducing the risk of hallucinations.

Latest Patents

Segal is named on one patent, 1 patent: “Guided Dynamic Construction of Large Language Model (LLM) Prompts.” The invention describes using multiple machine-learning models, each trained to identify a specific topic in text, to generate topic-specific prompts from an email thread. A user’s responses to those prompts are combined with customer relationship management data and enterprise information, such as calendar details, to construct a prompt for an LLM. The model then uses that context to draft business correspondence, such as an email reply. The approach is designed to improve relevance and reduce unsupported or inaccurate content.

Career Highlights

Segal’s listed patent reflects work at the intersection of machine learning, business communication, and enterprise data. It describes a guided process that brings user input and organizational context into LLM-generated correspondence.

Collaborations

Segal’s listed coworkers include Roy Eisenstadt and Abed El Kader Asi. The patent’s described system also brings together machine-learning models, user-provided responses, CRM information, and enterprise-suite data.

Conclusion

Bar Segal’s patent presents a structured approach to making LLM-generated business replies more context-aware and dependable by grounding them in topic-specific user input and enterprise information.