Are you tired of spending countless hours sifting through mountains of documents to find the insights necessary to make complex corporate decisions? As businesses continue to grow, so does the volume of unstructured documents, and processes, making it increasingly difficult to uncover actionable insights. If so, it’s time to harness the power of Generative AI.
In this talk, Arun will share his real-life experience building a Finance Copilot, a Generative AI solution that acts as an intelligent copilot for the finance team. The solution has helped thousands of users save countless hours and significantly impact the bottom line by providing answers to complex financial questions in seconds. But building a Generative AI solution has its challenges.
In this thrilling Generative AI session, we will enable the audience to harness the power of Generative AI, by diving into the intriguing world of question-answering systems & conversational agents designed to communicate with sensitive internal and external financial documents, complex process diagrams (flow charts) with several interconnections, accounting manuals. We will explore the potential of AI-driven question-answering systems that not only decode and interpret intricate process diagram interactions but also communicate effectively with various data sources depending on the nature of the question being asked along with several natural language processing & generation techniques.
The audience will take away the insights of real-world experience in building a Finance Copilot.
- How to mitigate the risks of exposing sensitive enterprise data when working with non-open-source Generative AI models like ChatGPT and GPT-4?
- “Talk-to-Flow” – How to communicate with complex process diagrams (flow charts) with several interconnections? How to pre-process the data to get it ready for GenAI?
- How to build a question-answering system that will dynamically talk to multiple data sources depending on the nature and complexity of the questions?
- How custom tools, and agents coupled with multiple advanced prompt engineering techniques like Prompt Chaining, Automatic Multi step reasoning and tool (ART), Tree of Thought (ToT), Retrieval Augmented Generation (RAG) etc, with prompt templates help in improving the quality of output?
- How to add citations & relevant facts to improve the credibility of the AI output?
- How to prevent hallucinations in Large Language Model (LLM)?
- What practical technical challenges are faced in building the Finance Copilot GenAI solution?
Acquire practical experience with hands-on exercises that help you apply the concepts learned from real-world experiences. Envision the potential of AI-driven conversation in transforming information exchange, decision-making, and knowledge transfer across various domains. Stay ahead of the curve by grasping the innovative capabilities of conversational AI and the future directions of Generative AI research and technology.”
