Avinaash Anand@avinaashanandk97

CompanyMatch addresses the following problem: A. Traditional company research and competitor analysis are extremely time-intensive processes. Investment professionals and business developers often spend days or weeks manually gathering information, which drastically reduces their productivity. B. The sheer volume of data available about companies and markets can be overwhelming. Sifting through this information to find relevant insights is challenging and error-prone and at times could lead to information being overlooked. CompanyMatch leverages IBM Watson (watsonx) and Exa to automate the discovery of similar companies, which can significantly streamline research for investment and business development professionals. Here are the key highlights: A. Programatic searching: Exa is used to search for information on companies and aggregate market data in target geographies. B. Intelligent data processing: Watsonx is employed at various stages to summarize search results and extract relevant information in a structured format. This includes summarizing company information, generating search terms for comparable companies, and extracting top comparable based on context from search results. C. Similarity analysis: Used cohere's embedding models to assess the semantic similarity of the identified comparable companies to the source company provided for a more nuanced comparison. On the technical front: A. XML prompting techniques: The flow utilizes XML prompting techniques to obtain structured output in a JSON format from watsonx models, enhancing the consistency and usability of the generated information. B. Few-shot prompting: Allows watsonx granite models to be more grounded in executing the task By automating these time consuming research tasks, CompanyMatch significantly enhances productivity, allowing professionals to hit the ground running when it comes to deal related due diligence or business development activities.