How I built a 30-minute research routine using Claude

Most reserch sessions don’t fail because of missing information..

They fail because there’s no structure to the time.

You open a tab, read half an earnings summary, get distracted by a headline, paste something into Claude without a clear question and end up 90 minutes later with a folder full of notes and no actual conclusion.

I’ve been there. The fix wasn’t finding a better tool.

It was building a tighter ruotine.

This is the 30-minute process I run when I need to evaluate a company quickly, decide if it’s worth going deeper and walk away with something I can actually use.

There’s a version of research that takes days. Deep dives, full earnings transcript analysis, multi-year financial modeling. That work matters, and AI handles it well.

But that’s not what most research sessions actually are.

Most of the time you’re answering a simpler question: is this worth my attention?

A company comes up in a screen. A reader sends a question. You see a name mentioned in a sector report. You have 30 minutes before your next call. Is this business interesting enough to go deeper?

That’s the question this routine is designed to answer.

Before I start the timer, two things:

One, I open a blank Claude conversation and paste in my research template at the top. Having the prompts ready means I’m not writing questions from scratch mid-session. I’m just filling in the company name and running.

Two, I open the company’s investor relations page in a separate tab. I want the most recent earnings release and the latest annual report within reach. I don’t read them before starting. I just want them there if I need to paste something specific.

That’s it. No other preparation.

The first prompt I run is the company overview. Not a deep dive. Just enough to understand what the business actually does and how it makes money.

Here’s what I paste in:

Give me a concise professional overview of [company name]. Cover: what the business does and how it generates revenue, the industry it operates in and the main competitors, how the company is positioned relative to those competitors, and any major developments in the last 12 months I should know about.

I read the output once, quickly. I’m not taking notes yet. I’m building the mental model.

By the end of this step I can answer: what does this business do, who does it compete with, and does the basic setup seem interesting?

If the answer to that last question is no, I stop here. Time saved.

If the business passes the initial filter, I move to the second prompt. This is where I try to understand whether the business has any durable advantage or whether it’s competing on price and hoping for the best.

Analyze the competitive position of [company name] from a long-term investor perspective. Evaluate: what advantages the company has over competitors, whether those advantages are structural or temporary, what could erode the position over the next five years, and how the financial performance reflects the quality of the business model.

This prompt does two things at once. It forces a moat assessment and it connects that assessment to actual numbers. A wide moat that doesn’t show up in margins or returns on capital usually isn’t as wide as it looks.

What I’m looking for: pricing power, switching costs, scale advantages, something that makes customers stay. If none of those things are present, the business might still be cheap enough to be interesting, but I need to know that going in.

One prompt, focused entirely on what could go wrong.

What are the three to five most important risks facing [company name] over the next two to three years? For each risk, explain why it matters, how likely it is to affect the business and what I should watch for in the company’s reporting that would signal the risk is becoming more serious.

I do this as a separate step rather than bundling it into the overview because risks tend to get buried when they’re listed at the end of a long analysis. Giving them their own prompt forces the output to treat them seriously.

By the end of this step I have a clear picture of the bear case. That matters. If I can’t articulate the bear case, I don’t understand the investment.

The last step isn’t another research prompt. It’s a synthesis prompt.

Based on what you know about [company name], give me a one-paragraph investment summary covering: the core thesis if you were considering a long position, the one or two things that would need to be true for that thesis to work and the main thing that could make it fail.

This is the step most people skip. They gather information and then stop. The synthesis is what turns information into a position.

The output from this prompt becomes my note for the company. If I come back to it in three weeks, this paragraph tells me exactly where I landed and why.

It is not a substitute for deep research on a position you’re actually building.

Before adding a company to a real portfolio, you want the full earnings analysis, the multi-year financial model, the management quality assessment, the sector comparison. That work takes much longer and it should.

This routine is for the first question: is this worth the deeper work?

Most companies aren’t. The 30-minute routine tells you that quickly so you can move on. When one is, it gives you enough of a foundation to know where to focus the deeper session.

For easy use, here are the three research prompts and one synthesis prompt in sequence:

Prompt 1, orientation: Give me a concise professional overview of [company name]. Cover: what the business does and how it generates revenue, the industry it operates in and the main competitors, how the company is positioned relative to those competitors and any major developments in the last 12 months I should know about.

Prompt 2, quality check: Analyze the competitive position of [company name] from a long-term investor perspective. Evaluate: what advantages the company has over competitors, whether those advantages are structural or temporary, what could erode the position over the next five years and how the financial performance reflects the quality of the business model.

Prompt 3, risk check: What are the three to five most important risks facing [company name] over the next two to three years? For each risk, explain why it matters, how likely it is to affect the business and what I should watch for in the company’s reporting that would signal the risk is becoming more serious.

Prompt 4, synthesis: Based on what you know about [company name], give me a one-paragraph investment summary covering: the core thesis if you were considering a long position, the one or two things that would need to be true for that thesis to work and the main thing that could make it fail.

Without a routine, every session starts from scratch. You’re deciding what to look at, in what order, with what questions. That decision-making overhead is where most of the time goes.

With a routine, you walk in with a plan. The first 10 minutes answer one question. The next 10 answer another. The last 10 produce something you can use.

Four prompts, 30 minutes, one paragraph you can keep.

That’s the whole thing.

The routine above gets you to a clear first answer fast.

But there’s a second layer I run when a company passes the 30-minute filter and I want to decide whether to actually build a position.

It adds roughly 20 minutes to the session and covers three things the quick routine deliberately skips.

These steps are what separate a watchlist company from one I’m willing to own.

I’ve written them up in full below, with the exact prompts and what to look for in each output.