Produce investment memos with Claude Cowork
In 15 minutes to a full investment memo. Task: The author used Claude Cowork to automate an equity research workflow for Adyen, involving data collection from investor relations and yfinance, building an Excel model with DCF and comps tabs, and drafting an investment memo. The process included setting up a folder structure, generating a business overview, and creating a scheduled task for weekly updates. Human involvement was limited to providing prompts and reviewing outputs, with the author explicitly stating they did not cross-check data against annual reports.; tools: Claude Cowork. Human role: The original post does not say how a person stayed involved.; limitations: The source is a secondary blog post with no independent verification of the financial model's accuracy or the tool's reliability. The author admits to skipping manual data cross-checks, relying on Claude's approximations. Scheduled tasks require the computer to be on and Claude active, limiting true automation. The '15 minutes' claim is anecdotal and may vary based on data availability and network speed..
Tools
Published on Agent Work Atlas:
Overview
In 15 minutes to a full investment memo
Workflow
The author used Claude Cowork to automate an equity research workflow for Adyen, involving data collection from investor relations and yfinance, building an Excel model with DCF and comps tabs, and drafting an investment memo. The process included setting up a folder structure, generating a business overview, and creating a scheduled task for weekly updates. Human involvement was limited to providing prompts and reviewing outputs, with the author explicitly stating they did not cross-check data against annual reports.
Reported result
The workflow produced a DCF model with three scenarios (Bear, Base, Bull) using specific assumptions like 9% WACC and 3% terminal growth, alongside a bullish investment thesis. The author reported the process took under 15 minutes but noted that data fetching took 5 minutes due to transcript availability issues. No independent verification of the financial accuracy or the final investment decision's success is provided in the source.
Human role
The original post does not say how a person stayed involved.
Limitations
The source is a secondary blog post with no independent verification of the financial model's accuracy or the tool's reliability. The author admits to skipping manual data cross-checks, relying on Claude's approximations. Scheduled tasks require the computer to be on and Claude active, limiting true automation. The '15 minutes' claim is anecdotal and may vary based on data availability and network speed.
Try it
Act as an equity research analyst using Claude Cowork. I will provide a target company ticker [replaceable input: e.g., ADYEY] and a local folder path [replaceable input: e.g., /research/ADYEY]. First, create a folder structure with subfolders for filings, transcripts, charts, model, and memo. Then, prompt Claude to download the last 5 years of annual reports and the last 4 earnings transcripts from the company's investor relations page, saving them to the correct folders. Use yfinance to pull 5 years of financial data and save as CSVs in the model folder. Next, instruct Claude to build an Excel model with a DCF tab using three scenarios: Bear (10% revenue CAGR, flat EBITDA margin), Base (18% CAGR, margin expanding to 55%), and Bull (25% CAGR, margin to 60%), assuming 22% tax, 9% WACC, and 3% terminal growth. Add a Comps tab with peer multiples. Finally, generate a 2-page investment thesis in a .docx file. Before finalizing, I must review the Excel formulas and the thesis for logical consistency. Do not publish or act on the thesis without my explicit approval. If data is missing, flag it in a README file instead of guessing.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.