From ChatGPT to Bloomberg AI
Claude and ChatGPT: handle reasoning and drafting, summarizing filings, checking the logic in a model, and explaining an unfamiliar concept. They compress research time but still need output checked against primary sources, since both can make factual or arithmetic errors.
Microsoft Copilot: automates the mechanical side of Excel work, building formulas, flagging inconsistencies, and speeding up formatting. It does not replace understanding modeling logic. Interviewers still test that directly.
Bloomberg Terminal AI features: support natural-language querying of market data and news. Access usually runs through a university trading lab or terminal license rather than a personal subscription.
Python paired with an AI coding assistant: such as GitHub Copilot or Claude Code, opens the door to quant modeling, backtesting, and automation. It helps students who are not computer science majors get productive faster, though it cannot substitute for understanding the statistics underneath the code.
AlphaSense: It is a research search engine used across investment banks and hedge funds to scan filings, transcripts, and broker research in seconds rather than hours. Many finance programs now offer student access through career centers.
Rogo AI: It supports deal execution inside banks: screening comparables, pulling filing data, and drafting preliminary models. Most students meet it for the first time during an internship rather than in class.
Tableau’s AI layer: It is built on Salesforce’s Einstein features. It turns raw data into dashboards through natural-language queries. The value here is communication, not computation. Analysis only matters once someone else can read it.
Zest AI: It applies machine learning to credit risk, moving underwriting beyond traditional FICO-based scoring. For students headed toward risk or fintech roles, it is a useful case study in how AI intersects with regulation and fair-lending compliance.
Wolfram Alpha: checks the math behind everything above: statistics, calculus, and optimization problems from portfolio theory and options pricing courses. It is not generative AI, but its computational engine catches errors before they reach a model.
