Monsaic Methodology: How We Structure and Source AI Stock Research

Why methodology matters

Anyone can point a language model at a ticker and get a confident-sounding paragraph back. That paragraph is not research — it’s a guess dressed in the vocabulary of research. The gap between the two is structure: a fixed set of questions asked every time, in the same order, with every factual claim traceable to where it came from.

Without that structure, an AI-generated stock writeup optimizes for sounding plausible, not for being checkable. It can drift toward whatever narrative the prompt implies, skip the inconvenient risk, or restate a stale number as current. A reader has no way to tell the difference between a well-supported claim and a fluent one.

Monsaic’s methodology exists to close that gap. The report format is fixed, the sourcing discipline is fixed, and the conditions that would invalidate the thesis are written down before the grade is assigned — not adjusted afterward to fit the conclusion.

The report structure

Every Monsaic report works through the same fifteen sections, in the same order, regardless of the stock:

1. 01Executive summary
2. 02Fundamental analysis
3. 03Balance sheet analysis
4. 04Catalyst analysis
5. 05Valuation & scenarios
6. 06Comparative precedents
7. 07Asymmetric risk profile
8. 08Risk assessment
9. 09Technical analysis
10. 10Insider & institutional signals
11. 11News sentiment
12. 12Management & governance
13. 13Narrative vs. substance
14. 14Dilution & shareholder value
15. 15Revenue quality

The order is not arbitrary. It moves from what the business actually does and earns, into what could change that (catalysts and valuation), into what could go wrong (risk, technicals, governance forensics), and closes on the two checks that catch a good story covering for weak fundamentals: whether the narrative matches the numbers, and whether shareholders are quietly being diluted.

Reading fifteen sections is not the point for most people — understanding the thesis, the risks, and the valuation is. The fifteen-section structure is what makes that summary trustworthy: it exists so nothing material gets skipped on the way to the verdict.

Sourcing and claim-checking

A claim without a source is an opinion. Monsaic reports carry a source list, and material claims trace back to where they came from — a filing, an earnings call, a disclosed financial statement, reported market data.

Not all sources carry equal weight. Primary filings and company-disclosed numbers are treated as the baseline; commentary, secondary summaries, and narrative framing sit below that and are used to explain context, not to establish facts. When a claim can’t be traced to something checkable, that’s a limitation worth stating rather than a gap worth papering over.

This discipline is slower and less impressive-sounding than an unsourced paragraph. It’s also the only version of AI stock research that a skeptical reader can actually verify.

Valuation scenarios

Monsaic does not reduce a stock to a single price target. Each report lays out four scenarios — bull, base, bear, and tail — and for each one specifies a price target, a probability, what would have to be true for that scenario to play out, and what would break it.

A single price target implies false precision: it suggests the future is one number away from being known. A scenario range does the opposite — it forces the analysis to state its assumptions explicitly, so a reader can see exactly which conditions the bull case depends on and exactly what would have to fail for the bear case to be right instead. The scenarios are not a hedge; they’re the actual shape of the uncertainty.

Kill criteria

Every Monsaic report defines its kill criteria — the specific, checkable conditions that would break the investment thesis — before the grade is assigned, not after.

This ordering matters more than it sounds like it should. A thesis that can’t be proven wrong isn’t a thesis; it’s a story that survives any outcome. Kill criteria are what make a Monsaic verdict falsifiable: concrete, observable events that would mean the original call was wrong, stated in advance so they can’t be quietly redefined later to match whatever actually happened.

In a recent Monsaic analysis of a major semiconductor supplier, the kill criteria included two consecutive quarters of hyperscaler capex plans rolling over, gross margin structurally falling below a stated floor without a clear mix-shift explanation, and competing silicon taking enough share to flatten the company’s core growth line. Each one is something a reader could check against a future earnings call — not a vague warning that “risks exist,” but a specific bar the thesis has to clear.

Grades and the final verdict

Alongside the narrative sections, Monsaic assigns forensic grades on management and governance, legal and regulatory standing, dilution, and revenue quality — each on an A-plus-to-F scale with a written summary explaining the grade — plus a separate one-to-ten score on how well the company’s narrative matches its actual financial substance.

These grades are diagnostic, not decorative. They surface the parts of a company that a pure growth story tends to gloss over: is management diluting shareholders to fund the story, is revenue quality holding up under scrutiny, does the narrative investors are told match what the numbers actually show. The report’s overall verdict — the call and its conviction level — sits on top of these grades, the scenario analysis, and the kill criteria together. A grade states what it found; it does not claim certainty about what happens next.

The simple read and the full report

Every Monsaic report exists in two forms built from the same underlying analysis: a plain-English simple read that preserves the full verdict, the thesis, and the key risks without requiring familiarity with equity-research vocabulary, and the complete fifteen-section report for readers who want the underlying depth.

The simple read is not a watered-down summary that drops the hard parts. It’s a design constraint: whatever conclusion the full report reaches, the simple read has to reach the same conclusion, just in plainer language. Reading five minutes of the simple read and reading the full fifteen sections should never leave a reader with different takeaways about the thesis.

Freshness and updates

Markets move, and equity research decays. Every Monsaic report carries an explicit “as of” analysis date and a price-as-of timestamp, so a reader always knows how current the underlying numbers are rather than assuming a report is live simply because it’s on the screen.

Price and market data in a report are delayed rather than live: during trading hours quotes are typically up to about 15 minutes behind, and outside market hours a report carries the most recent closing price. Every report stamps the exact time its price was observed, so the reader always knows which number the analysis was built on.

A dated report is not a live report. Freshness means the report states its own age honestly and can be regenerated on demand when a reader needs a current read — it does not mean every report is perpetually up to the second. Treat the “as of” date as load-bearing information, not fine print.

Limitations

Monsaic reports are educational research and analysis, not personalized financial advice. They do not account for an individual reader’s portfolio, risk tolerance, tax situation, or financial goals, and they should be one input into a broader research process, not the final word.

Report quality depends on the quality and availability of the underlying public sources — filings, disclosures, market data. When source material is incomplete, delayed, or itself unreliable, the resulting analysis inherits those limits. No report guarantees future accuracy, and no grade or scenario probability is a promise about what a stock will actually do.

Hold a real report against this page

The test of a methodology is the output it produces. Open a covered stock’s excerpt and hold it against this page — the verdict, conviction, thesis, and the condition that would break the case are its first products, and the full report carries the valuation scenarios and sourced claims described here.

Browse covered stocksSee pricing

Keep reading

• AI stock research, done rigorously — the Monsaic pillar
• How Monsaic ranks its sources
• How Monsaic checks claims
• Generate your first Monsaic