My favorite AI earnings diligence technique (steal my agent for free)

Many large companies today operate complex global logistics - but investment teams typically are structured around regional experts.

London covers EMEA tickers, HK / SG for Asia, and NYC for US.

This is a major gap especially for physical goods tickers.

For example: If you're researching Nike ($NKE), how often are you reading (and understanding) its suppliers' filings in Asia?

These suppliers can help indicate downstream margins, demand, and other color commentary before $NKE's own earnings.

Unfortunately, most investors are not digging into these reports.

The problem is following those connections across exchanges and reporting calendars, especially when filings are in Chinese, Korean, or Japanese.

So this week, I'm going to show you how to use AI to look for earnings signals across a company's supply chain.

We'll be using Garmin $GRMN as the example.

Written walkthrough below but I also posted a video essay on this too for you to follow along:

​

​

The Prompt

You can do this analysis on any platform of your choice with the necessary MCP servers connected. Those connections need to provide access to international earnings transcripts, filings, and other relevant source documents.

I used AlphaSense for this one as it already has the research content connected inside its platform.

Their team was nice enough to give our subscribers a free one-month trial to try this prompt.

You can download my agent prompt here.

I set it up as an agent in AlphaSense, then asked it to run on Garmin.

Once you've loaded the prompt, give it your target company or ticker. Customize the reporting period and earnings questions you want answered.

The Result

Full output here: Garmin_supplier_analysis.pdf​

In the results, AI not only identified Garmin's key suppliers but also indicated its key upstream suppliers reported record growth few days before earnings.

That gives me a research checklist: which companies to follow, what changed, and which earnings assumptions need another look.

In this case, Garmin ended up beating last quarter.

So how would I generalize this technique?

First, I'd start with physical-goods companies whose suppliers and sales partners publish useful financial updates. Earlier reporting and meaningful exposure to the brand make those disclosures more valuable.

Nike and adidas are good examples to investigate. The same approach can fit consumer electronics, autos, and appliances.

Here are the questions I'd want answered:

• Orders: Are manufacturers seeing stronger bookings, cancellations, or shorter lead times?
• Costs: Are suppliers raising prices, and when could those increases reach the brand?
• Inventory: Are retailers selling products to consumers, or building stock that may need discounting?

An increase in factory orders could support a stronger revenue outlook. Higher input costs could hit margins later, depending on contracts and inventory.

That distinction matters when you're building an earnings view. Sales growth and margin pressure can happen together.

International coverage gives you more places to look for these clues. AI can help you work through documents scattered across countries and languages.

Finally ... a useful signal still needs to be compared with consensus. Improving demand alone doesn't establish an earnings beat.

Try this out and let me know what you think!

Personal

I'm back on the road for the next three months: the Middle East first, then Australia, followed by Singapore and Hong Kong.

I may also be hosting finance AI buildathons and teaching sessions with GLG along the way. If you're in any of those regions, hit reply and let me know!