Hertz ($HTZ) surged 125% last week on news of Bill Ackman’s Pershing Square taking an activist position in the company.
Naturally, I asked myself: “How can I use AI to profit off of event driven trades like this?”
$HTZ surged 125% between Pershing Square’s activist SEC filing and Bill Ackman’s X post …. which accounts reported the raw docs before Bill officially wrote about it?
One of my best friends from high school makes a living having a fast trigger being the first to trade news events like this…with X scraping as a tool in his arsenal.
Inspired by this, I created this gameplan:
Use ChatGPT o3 to identify exact posting time of the SEC filing (this model has great web search capabilities and reasoning)
Use Grok to identify which account(s) FIRST did the work to analyze the numbers and post about it
Tag these accounts as “high alpha” for news based trading in the future (most accounts just repost news, I want to find the original high signal sources)
I’ll show you below my raw prompts you can steal + my raw GPT interactions + the results on which account AI identified as the best for this trade 👇
The Prompt Chain:
Prompt 1 (Use ChatGPT o3 for best web search reasoning)
I want to confirm the official public disclosure date and time of Pershing Square’s stake in Hertz by searching the SEC EDGAR database (or its RSS feeds) (https://www.sec.gov/search-filings) for the relevant filing, likely a Form 13F-HR or similar, expected around April 16, 2025. Use Pershing Square’s CIK to locate the filing and determine the exact acceptance or filing timestamp. The Edgar website is https://www.sec.gov/search-filings.
I’m a trader looking to identify the first X account that reported Pershing Square or Bill Ackman taking an activist position in Hertz. The news was first filed to the SEC on April 16, 2025, at 07:43:09 AM ET, so the first X post about this event after 07:43:09 AM ET on April 16, 2025, wins.
Search X for the first post on X reporting this news after the SEC filing, using a broad range of keywords to capture variations in phrasing. Include terms like ‘Pershing Square Hertz,’ ‘Bill Ackman Hertz,’ ‘Pershing Hertz,’ ‘Pershing $HTZ,’ ‘SEC filing Hertz Pershing Square,’ ‘Pershing Square Capital Management Hertz,’ ‘Pershing Square Capital Mgt $HTZ,’ ‘Hertz stake Pershing,’ ‘Ackman stake Hertz,’ ‘Hertz shares Pershing,’ ‘HTZ Pershing Square,’ and other combinations that might appear in financial posts. Ensure the search captures posts from any account—individual, anonymous, or financial, including the smallest accounts with minimal followers—since I need the fastest source to trade on such news.
To narrow the search window, use advanced search operators like ‘since:2025-04-16 until:2025-04-17’ and filter for posts specifically between 07:43:09 AM ET and 09:00 AM ET on April 16, 2025, to focus on the immediate aftermath of the filing. Sort results by timestamp to identify the earliest post.
Cross-check the timing of the identified post with stock price movements to ensure it aligns with pre-market activity.
This was 10 minutes faster than Bloomberg (Bloomberg Law posted ~8:01) and 17 minutes faster than Zerohedge.
If you trade premarkets, this speed edge could have given you double digit returns by front running Bloomberg (premarkets settled +19% at open and @zohmbastic identified this <10% up). I’ve given this account a follow as a “high signal account” on my X list 🙂
What You Can Learn for Future Prompting
Specify Criteria Early: Mentioning all account types upfront ensures the AI searches broadly from the start, avoiding the need for follow ups.
Include Validation: Asking for a cross-check (e.g., stock price movements in Prompt 2) ensures the result is reliable without needing additional prompts.
Provide Enough Detail: Including the keywords in Prompt 1 gives the AI enough to work with, reducing ambiguity.
