Today, I was amazed by the demo video of #Claude AI from @AnthropicAI. As a professional investor focusing on US stocks and #Web3 projects, I have to keep an eye on hundreds of companies every year, especially during earnings season, which is like a battlefield!

The practical case shown in the video is quite common: I just sat down in the office at 2 PM, and the PM (Investment Director) rushed in urgently:

“Sarah, the company we are focusing on, Velocity Athletic, just released its earnings report, and revenue plummeted by 12%, but the stock price actually rose by 17%, reaching $71. You need to figure out whether this surge is due to favorable strategies or just a false rebound. We need to make a decision before the market closes today: should we sell to take profits or continue to hold?”

This logic is similar to the non-farm payroll data this Friday + the revised employment data for the last two months (May and June were revised down by 260,000 jobs, equivalent to the entire population of Miami suddenly becoming unemployed). In such situations, you need to make a judgment call immediately and provide feedback to the US stock market, hedging with crypto assets. This requires instant response and data support, which can effectively avoid the weekend's sell-off.

Let’s look at the normal path for a financial analyst (estimated 4-5 hours):

1️⃣ Go to S&P Global to pull earnings reports and meeting minutes.

2️⃣ Open FactSet / Morningstar to check historical data and valuations.

3️⃣ Check Edgar for SEC announcements.

4️⃣ Dig through internal memos and research reports.

5️⃣ Then create tables, draw charts, and write memos.

This entire process takes at least 4-5 hours, manually switching between several systems, with inconsistent information that can easily lead to errors.

💡 This is when #Claude comes in, truly like having a super analyst assistant:

I simply give #Claude the command: “Help me analyze the earnings performance of Velocity Athletic, determine if the stock price movement is reasonable, and produce an investment memo.”

🔍 1️⃣ Multi-platform data retrieval + semantic understanding

Claude doesn’t just scrape data; it understands the context I need:

• It pulls the conference call content from S&P Global and extracts key information, such as the CFO mentioning in the Q&A, “Due to tariffs, the profit margin decreased by 400 basis points”;

• It retrieves historical data and competitor comparisons from Morningstar;

• It finds the past analysis framework for Velocity from our internal documents (box);

• It pulls financial data for 8 quarters from Delupa and directly creates trend charts;

• All content is cited with sources, allowing for easy double-checking, greatly enhancing compliance and trust.

📊 2️⃣ Generate visual charts + core models

I ask #Claude: “Help me create an annotated price chart, a peer valuation comparison table, and a DCF model.”

It gradually accomplishes:

• An annotated stock price chart showing events (like CFO selling shares, emergency board meetings, earnings release times, etc.);

• A clear comps table: Velocity’s EV/EBITDA is 21 times, while the industry average is only 16 times, indicating overvaluation;

• A complete DCF model: clear assumptions, easy scenario switching, and even the WACC calculation is standard;

• The reasonable stock price derived from the model: $54, far below the current $71, indicating that the current rise is overly optimistic and has significant bubble components.

📝 3️⃣ Automatically generate institutional-level investment memo

Finally, I say: “Use our fund's internal template to produce a memo for the PM’s reference.”

Claude directly calls the internal memo template and automatically fills in the following content:

Recommendation conclusion: Suggest selling to take profits at this high point and wait for a price correction to re-enter;

Logical support: Quantitative support from financial fundamentals, market reactions, industry comparisons, and valuation deviations;

Potential risks: For example, if tariffs ease in the future or if there are other reasons for the CFO's selling, it may lead to further stock price fluctuations;

Clear citations: All data includes original source links for PM or legal review.

After watching the entire demo video, I feel that #Claude is truly the “#AI nuclear weapon” for earnings analysis. Currently, the API data interfaces supporting crypto data are still relatively few, but if improved in the future, this could become a killer tool for #Web3 investment research.

If we used to analyze projects with “ten fingers flying + nine screens checking data + five heads working together,” now with #Claude, it’s like having a gold medal analyst who never goes offline, doesn’t complain about overtime, and has skyrocketing efficiency📈!

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。