Damn it, there are a bunch of financial service applications that need to take a break. OpenAI has launched ChatGPT for Financial Services, which directly targets investment banks and stock research, bringing analysts' daily work into one workspace. There are several practical functions: financial data such as Daloopa, PitchBook, LSEG, Crunchbase, Quartr, etc. can directly enter the research process, and FactSet, S&P Global, Preqin, Datasite that institutions have already purchased can also continue to receive them. Analysts can allow it to search for companies, pull historical financial data, read financial reports and conference calls, conduct comparable company analysis, explain EBITDA adjustments, and then directly enter Excel modeling, update models, and conduct scenario analysis. Finally, according to the bank's own Word, Excel, and PPT templates, the research results can be transformed into memos, valuation materials, and workbooks. Another important feature is' traceability '. Financial institutions are most afraid of AI making up numbers in a serious manner, so OpenAI makes references to specific data and original text, and research results can be traced all the way to financial reports, tables, and documents. The output in Excel should also be kept within the original formulas and workbook structure as much as possible. In addition to enterprise functions such as permissions, encryption, and audit logs, it solves whether AI can truly enter the production environment of investment banks. Its true value, in other words, is to compress the process of a junior analyst moving back and forth between Bloomberg/FactSet, Excel, Word, PPT, and various databases. In the past, a financial AI startup could sell a product by doing AI to read financial reports, another product by doing AI modeling, and another product by doing AI research reports. Now OpenAI is starting to directly integrate data, reasoning Excel、 Package documents and enterprise permissions together. In the future, if financial AI applications want to survive, it will become increasingly difficult to just do GPT+financial data+UI. The ones who can still charge are either those who have data that others cannot access, or those who have been deeply involved in trading, risk control, compliance, and execution processes. OpenAI has directly adopted the workflow of financial institutions this time.