IASSIST 2025: IASSIST at 50! Bridging oceans, harbouring data & anchoring the future


Identifier Detective: Using AI to find information on Database Identifiers

As business librarians, we often get these questions, "What is an Identifier – is it unique to companies?"; “What identifiers are in capital IQ ?"; "How is CUSIP different than CIK?"; "Can you help me match my dataset to get industry NAICS codes?". I created a chatbot called "Identifier Detective" that helps users get answers for these questions interactively. To create this chatbot, I use a spreadsheet that matches different financial databases such as Capital IQ, Bloomberg, LSEG workspace with corresponding identifiers such as CUSIP, GVKey, ISIN, SEDOL, CIK, Duns Number, Ticker. Users can customize their questions with the chat bot. The chat bot creates a folder with user’s name and saves their answers in their own folder. If the user wants to find which databases have GVKEY and CIK – the chatbot will use the spreadsheet to provide the answer Capital IQ. It allows the user to upload their own dataset in their folder. For e.g. the user has data from Compustat and wants to know how to match the companies to get CIKs. Then, the chatbot will tell them their identifier is from Compustat and they can use GVKEY to match with Capital IQ and get CIKs for the companies. The chatbot provides more information on the Identifiers. For e.g. if users asks about CIK, it provides curated information about CIK that is in the spreadsheet. In this presentation, I will show how to create such a chatbot, show a live demo of the chatbot, limitations, and lessons learned when creating such chatbot.

Inku Subedi
Harvard Business School
United States