The world of business and law is characterized by intricate contracts and documents that demand close attention to ensure compliance with various laws and regulations, as well as proper filing and updating. Tracking legal cases and outcomes is a full-time task. Some of these documents can be hundreds of pages long and require hours of review. Mayo Oshin, an artificial intelligence engineer based in the UK, aims to revolutionize this process using AI. His latest project, Warren Buffett, focuses on the financial sector and its documents.

Oshin named the bot after Warren Buffett, the famed value investor, as it analyzes financial documents in his style. Oshin is based in London and contributes to LangChain, an open-source framework that offers several tools and resources to aid AI application development. According to Oshin, there is a growing trend among companies to incorporate the concept of retrieval in their AI systems. Retrieval refers to the ability to chat with data, which has become an essential requirement for many businesses.

Having experimented with retrieval capabilities for some time, Oshin observed a rising demand for practical examples of how this technology could be applied to various documents and data types. For instance, annual reports can be hundreds of pages long, and investors in Tesla may want to extract relevant information, such as current risk factors or management performance. The goal of the Warren Buffett project is to demonstrate how AI can help analyze complex and lengthy documents, extracting critical information quickly and easily using retrieval capabilities. Users can converse with data and obtain meaningful insights, which should lead to better decision-making.

The chatbot is capable of analyzing large and complex documents, enabling users to retrieve relevant sections of their document based on natural language, rather than having to read the entire document. Oshin designed the chatbot to demonstrate its potential beyond one-directional interactions, which have been the primary focus of AI discussions. The demo can perform an analysis over time using a time series approach. For example, it can analyze cash flow performance over several years to reveal trends.

Oshin believes that AI can be cost-effective, despite popular belief that it is expensive. He acknowledges that people are both excited about and scared of what AI means for knowledge work, a concern driving the interest and discussion around AI and its impact on the workforce. Oshin further adds that the people working on AI research themselves do not fully understand its capabilities, leading to uncertainty in the area.