5 AI Steps to Effortlessly Understand Any Business

…without reading the annual reports

Jul 05, 2025

Hello, Fellow Stock Pickers

We all know the Peter Lynch quote:

“The person who turns over the most rocks wins the game.”

Good news:
With AI, you can now turn over 10x more rocks…

Most investors waste hours clicking through reports, opening tabs, and hoping insight magically appears.
They start with the 10-K.
Scroll through headlines.
Maybe search for “margin” or “growth.”
Then get overwhelmed … and quit.

But research isn’t linear.
It’s iterative.

Your goal on the first pass isn’t to build the perfect investment thesis.
It’s to get a clear snapshot of the business.
Enough to know:

Do I go deeper, or do I move on?

That’s why I use his 5-step system.
It helps me go from “What is this company?” to “Here’s how it works + whether it’s worth more time” usually in under 60 minutes.

And today, I’m going to show you exactly how I used it on Semco Technologies , a French IPO in the industrial equipment space for semiconductors.

Let’s break it down in 5 steps:

Step 1:Start With a Solid Prompt, Then Adapt It to the business

Semco came up during a chat with an investor friend.

I was curious
But not "let me sink 3 hours into a 300-page PDF" curious.

So I used my system’s first move:
Drop the company into my standard research prompt.

Not a “pretend you’re Warren Buffett” kind of prompt.
A prompt structured around the questions I actually care about:

What does the company do?
Who are the customers?
What drives growth?
Is the revenue recurring or one-off?
Who are the competitors?
What’s the margin profile, ROCE, and capex intensity?
How cyclical is the business?

Now here’s the part most people skip:

Every time I hear about a new company, I give this base prompt to ChatGPT,
and I ask it to adapt the questions based on the company’s industry and business model.

It helps make the questions more focused so I don’t get vague answers and can go straight to what really matters.

Here’s my standard prompt (feel free to steal and tweak it):

MASTER PROMPT – DEEP BUSINESS ANALYSIS OF A PUBLIC COMPANY

You are a financial analyst specialized in deep business model analysis.
Your mission is to analyze a company as if you were acquiring 100% of the business, not just buying shares on the market.

I will give you the name of a company.
Your task is to:

1. Use the standard analysis framework below
2. Automatically enrich it with industry-specific and business-specific questions
3. Return a fully structured and detailed prompt, ready to use in DeepSearch or a GPT-based research tool

Your goal

Produce a comprehensive and actionable business model analysis, based on internal data (reports, filings, investor materials) using only reliable sources, including:

- Official documents: annual reports, IPO prospectus, investor presentations
- Competitor data: comparable companies, margin structures, strategic positioning
- Industry research: market reports (IDC, McKinsey, Gartner, Yole, etc.), technical white papers
- Expert insights: executive interviews, industry podcasts, analyst briefings, specialist newsletters
- High-signal sources: Seeking Alpha, TIKR, Stratechery, company transcripts, etc.

Use concrete examples wherever possible (client case studies, unit economics, cost structure, etc.).

Standard Framework to Enrich Automatically

1. Business Understanding

- What is the company’s exact activity?
- What does it sell? What core customer problems does it solve?
- Who are its typical customers? What is their buying process?
- Why do customers choose this company over its competitors (price, technology, reputation, integration, service, etc.)?
- What is the nature of revenue: recurring (SaaS, contracts), one-off (capex), hybrid?
- What is the typical gross margin structure?
- Is this a simple, predictable, project-based, or cyclical business?

2. Market and Competitive Position

- What is the real Total Addressable Market (TAM/SAM)?
- What are the main industry tailwinds or growth trends?
- Who are the direct and indirect competitors?
- What are this company’s competitive advantages (tech, cost, IP, brand, regulation, network effects)?
- Are these moats durable over 5–10 years? Why or why not?
- Any threats of disruption (tech shift, commoditization, regulatory changes)?
- Does the company have pricing power?

3. Financials

- Revenue, EBIT, and FCF evolution over the past 5–10 years (or forecasts if IPO)?
- Margin trends: gross, operating, and net margins?
- ROCE / ROIC – is capital allocation efficient?
- Is the company self-funding? Does it generate sustainable free cash flow?
- What is the debt profile? Any dilution history or capital dependency?
- How cyclical or seasonal is the business? Is visibility high or volatile?

4. Management and Strategy

- Who leads the company? Are they founders? What’s their ownership stake?
- Have they proven operational success in this industry before?
- What is their stated long-term strategy?
- How have they allocated capital historically (organic growth, M&A, dividends, buybacks)?

What You Must Deliver

Return a clean, enriched prompt, ready for DeepSearch or GPT, that asks:
- the right business-level questions
- queries all official documents
- integrates sector dynamics
- references peer comparisons
- and pulls insights from expert sources

to produce a comprehensive business model analysis, focused, concrete, and useful for high-conviction decision-making.

Now simply ask me:
"What company do you want to analyze?"
Then build the enriched, sector-aware prompt.

Step 2: Feed the Prompt Into Deepsearch, Then Let It Do the Heavy Lifting

If you only take one thing from this newsletter, my friend let it be this:

Use Deepsearch.
It’s like having a research intern that never sleeps.
Or better: an analyst you pay $20/month… and it actually does the job.

I use it every time I discover a new company.

Once my prompt was tailored to Semco’s business model,
I dropped it into google's gemini Deepsearch.
(With access to public filings, websites, and docs scraped from AMF.)

30 minutes later, I had:

A 22-page deep-dive
A clear business model summary
Management quotes
Strengths, weaknesses, and risks already organized

No endless scrolling.
No Ctrl+F’ing through PDFs.

Now it’s up to you:

You can read the full report.

You can ask AI to summarize it into a 10-line elevator pitch.

Here’s a quick snapshot AI gave me as an overview before digging deeper

Semco Technologies is a French semiconductor equipment company specialized in designing and manufacturing electrostatic chucks (eChucks) — devices used to hold silicon wafers during critical steps of chip manufacturing. Unlike diversified ceramic players, Semco is a pure player with full vertical integration and focus on this niche. eChucks are crucial for process stability and are used across devices like smartphones, EVs, and industrial electronics.
In 2024 (pro forma), Semco generated €26M in revenue, with 65% gross margin and ~37% EBIT margin, positioning it as a high-margin player. Its top 10 clients represent 93% of sales, showing strong customer relationships but also concentration risk. The IPO brings fresh capital, used to fund industrial automation, R&D, and commercial acceleration.
Management expects €55M+ revenue by 2028 and >40% EBIT margin, driven by demand for next-gen machines and process customization. Key risks include customer concentration, production capacity limits (single production line), and geopolitical exposure (US/China tensions). Despite that, the high margins, niche focus, and embedded client relationships make it a potential compounder—if execution and scaling succeed.

(My favorite? Step 3 where it all starts to click.)

Step 3: Go Deeper with a 10-Minute Audio Summary

After getting the initial overview, instead of reading it all, I just ask IA it to turn the content into an audio summary.

In a few seconds, I had a 10-minute podcast version:

Clear synthesis of what Semco does, how it makes money, and where the risks are
No hallucinations everything came from official filings
Perfect to listen while walking, or training

I used to spend 2–3 hours per company. Now I get clarity in 15 minutes max.

If you’re not turning long reports into audio, you’re missing an easy edge.

Step 4: Test the Bull Case

Now that I’ve got the facts it’s time to challenge the narrative.

Inside Gemini (or ChatGPT), I take the core of the Deepsearch report and run a “Bull vs Bear” simulation with this prompt:

“Give me the most convincing Bull Case and Bear Case for investing in Semco Technologies based only on data from the IPO doc and peer benchmarks. Be specific and objective.”

This step forces clarity:

If the Bull Case sounds vague, there’s probably no edge.
If the Bear Case hits hard, I look for ways to mitigate it (or I move on fast).
If both sides feel real, it’s worth deeper analysis.

Example output:

Bull Case: Strong margins, niche positioning, embedded relationships, structural growth in semiconductors.
Bear Case: Customer concentration, single production site, IPO mostly to cash out, limited float.

This step saves me from falling in love with a story too early.
It’s fast, brutal, and clears the fog.