Front office: Research and diligence
3분 읽기Analysts and deal teams work with Claude as the first reader across their coverage universe, including filings, transcripts, broker notes, expert calls, data rooms, and the firm’s own research history. Claude can update models, use connectors to pull the data it needs, and run scheduled tasks overnight.
Main use cases
Updating and building models in Excel
Analysts can ask Claude to roll a coverage model forward from the 10-Q, rebuild a broker’s hard-coded sheet as live formulas, or add a sensitivity tab. Deal teams can ask for leveraged-buyout models, credit models, or cap tables, built from confidential information memorandums (CIMs) and data rooms. Claude edits workbooks directly and cites the cells behind each answer.
Processing events overnight
On earnings night, a scheduled agent reads the release, transcript, and filings through connectors. It compares the results with consensus and with the team’s model, updates the model, and drafts a first-take note, so the analyst starts the day with a marked-up model and a note to edit. Anthropic’s financial services plugins are a starting point for these steps.
Researching the team’s archives
Because Claude Enterprise connects to internal document search, vendor data, and the research management system, teams get answers that cite multiple sources. Connectors that each person signs in to use that person’s existing permissions, so Claude reaches only the documents that person has access to. Private-markets teams ask Claude to get them up to speed on a new investment and to track portfolio performance.
Conducting deep research, diligence, and red-teaming
For open questions, such as sizing a market or assessing a founding team, an agent breaks the question into subquestions, searches independent sources for evidence, and returns a cited report or a diligence checklist with findings. Before a pitch or an investment-committee meeting, Claude can follow a red-team skill to argue the bear case against the team’s own thesis.