Eleven prompts, and a standard for judging what comes back. The first one treats your company and your portfolio as a single balance sheet, because if you own a business your largest position isn’t in your brokerage account. The rest cover any company or fund. Every prompt carries an instruction block that tells the AI to use current data, name its sources, admit what it couldn’t verify, and lead with the answer instead of burying it. Built on the same diligence habits behind An Outsider’s Playbook. Free. No email required.
Whose words are whose
The prompts here are mine. The answers are not. Everything these prompts produce comes from whichever AI assistant you paste them into. It is not my analysis, not my opinion on any security, and not a view I hold or endorse. I have never seen it. Two people running the same prompt on the same ticker on the same day can get different answers, and both can be wrong.
That is worth saying plainly, because a free tool on an author’s site invites the assumption that the author is behind the output. I am behind the questions. The questions are the part I have spent four decades refining, and they are the part the book is about.
The rule behind the first test has a name and a chapter. It is Chapter One of An Outsider’s Playbook: serve the cake first, and hand over the recipe only if someone asks. Every prompt here enforces it. Read Chapter One →
What a good answer looks like
Every prompt ships with an instruction block that sets a standard. Hold the output to it. If an answer fails these tests, it is a weak answer — push back on it or run it again rather than acting on it:
It leads with the verdict. The conclusion is in the first two lines, not the last paragraph. If you had to read to the bottom to find the point, the answer failed.
It dates its figures. Every number carries an as-of date. Undated numbers are unusable.
It names what it could not verify. Silence about sourcing is the tell. A good answer says which figures came from a filing and which did not.
It knows what it is looking at. A company and a fund are different questions. If the answer discusses management or insider activity for an index fund, it has hallucinated the premise.
It distinguishes analyst work from machine ratings. Many small companies carry no human analyst coverage at all, only algorithmic fair values presented in identical language.
It refuses when it should. An assistant that cannot search the web should say so and stop, not answer from memory.
Before you act on any of it
Verify every figure against a primary source — the company’s own filings, the fund’s prospectus, or exchange data. Treat the output as a structured starting point for your own work and for a conversation with a licensed professional. Never as a conclusion.
Start with what you already own
One position per line. Ticker, then shares, then average cost. Fills the first four prompts.
Start here if you own a business
01The Founder’s Balance Sheet
Open with the answer: does this portfolio hedge my exposure to my own company or quietly double down on it? One sentence, then the single change that would fix it. Before any analysis. I’m an entrepreneur. My largest asset is my own company, in [industry], operating primarily in [geography]. My largest customer is roughly [customer concentration] of revenue, and my equity in the business is [liquid / illiquid]. Here is my public-market portfolio: [paste holdings above]. Beneath the verdict, treat the business and the portfolio as one balance sheet and show the overlaps: same sector, same customers, same macro sensitivity, same geography, same interest-rate exposure. If I hold funds, look through to their holdings, because a broad index fund can easily contain my own customers or my competitors. Then tell me how the portfolio’s role should change given that my income, my equity and my time are already concentrated in one private bet. Close with the correlation I am least likely to have noticed.
02Portfolio X-Ray
Open with what this portfolio is implicitly betting on, in one sentence, and the single change that would most improve its resilience without gutting its upside. Before any analysis. Here is my portfolio: [paste holdings above]. I also hold [cash] in cash, and these assets outside the brokerage account: [outside assets]. Then the three findings that produced that verdict, one line each. Beneath them, analyze it as one machine rather than a list of tickers: concentration by position, by sector, and by underlying holding, looking through any funds to what they actually hold, because two funds can own the same companies and appear diversified while moving together. Identify the positions likely to move together, and state clearly whether that is a qualitative judgment or whether you actually examined return correlations. Name the single macro event that would hit everything I own at once. Close with the bet I am making that I probably did not intend.
03Add, Trim, or Hold
Open with the verdict table: every position, one line each, add, hold or trim, with your one-line reason. Before any discussion. Here is my portfolio: [paste holdings above]. My original reason for buying each: [reasons, where I gave them]. Then the two rankings: which positions most deserve new capital, and which are still here by inertia rather than conviction. Beneath that, expand only on the positions where your verdict is likely to surprise me, with the strongest argument for and the strongest argument against each. Where I haven’t given you a reason for buying, say so and do not invent one on my behalf. Close with the single position you would sell first if I had to raise cash tomorrow.
04Liquidity & Lumpy Income
Open with the liquidity ladder — what gets sold first, second and last — and the reserve number you think I should be holding. Before any explanation. I’m an entrepreneur with irregular income. Here is my portfolio: [paste holdings above]. My business burns roughly [$burn] a month and I currently hold [$reserve] in cash. Beneath the ladder, justify each rung: which positions I could sell quickly without material damage, which would be painful to exit, and the tax consequence of each step stated as we go. Account for settlement timing, mutual-fund redemption mechanics, and any early-withdrawal penalty on retirement accounts — a position inside a retirement account may belong last no matter how liquid the security itself is. Ask me for the account type of any holding where it would change the answer. Close with how many months of burn the ladder actually buys me before I would be forced to sell something I don’t want to.
One company, one fund, one decision
Works for a company, an ETF, or a mutual fund. Every prompt tells the AI to check which one it is first, and to switch to fund questions when the answer is a fund. Fills the last seven prompts.
05Is This the Right Vehicle?
For ETFs and mutual funds. The question almost nobody asks, and usually the one worth the most.
Open with the answer: should I keep [TICKER] or switch, and if switch, to which fund and what it saves me in dollars. One line, before anything else. Then the comparison of every materially cheaper or structurally better fund tracking the same index: total expense ratio; fund structure, including whether it is a unit investment trust or an open-ended fund and what drag that structure creates; which share classes exist and which of them a retail investor can actually buy; assets and trading liquidity; and capital-gains distribution history. Present it as a table. Beneath the table, tell me what index or strategy it tracks and how closely it has actually tracked it, and quantify the fee difference over [years] years on a [$amount] position. Close with the honest reasons a rational investor might still hold the more expensive one, and whether any apply to me.
06Diligence the Business
Give me your verdict on [TICKER] as a business in one sentence, then the Bull, Bear and Base case in one line each, before any analysis. Beneath that, descend into what produced them, in the order that matters most: how it actually makes money and how durable each revenue stream is; unit economics and margin structure; the real competitive moat and the specific mechanism by which it erodes; management’s capital-allocation record; balance sheet strength; and insider or management ownership plus their recent reported transactions, using whichever disclosure regime governs the market where it trades. If you cannot verify those transactions in the relevant register, say so rather than characterizing them, and say whether the register exists at all in that market. Flag anything that looks like financial engineering rather than operating performance. Close with the specific evidence that would prove each of your three cases wrong.
07Buy Decision
Lead with your read on [TICKER]: which of buy now, wait, or avoid has the strongest case given my constraints, in one line, followed by the single variable that would change that answer. Then the three findings that drove it, one line each. My horizon is [horizon], my risk tolerance is [risk], and this would become roughly [position %] of my portfolio. Look up the current price yourself. Beneath the call, expand: what is already priced in, valuation against its own history and against peers, the quality of the earnings behind the multiple, and which catalysts and risks the market is currently paying for. If earnings are negative or near zero, say so and use a metric that actually works instead of reporting a distorted one. Map the realistic downside, not just the upside, and give me the strongest case against your own call. Close with what would flip it.
08I Already Own It
Answer in your first line: is the reason I bought still true — confirmed, broken, or quietly replaced by a different one? Then give me the three exit paths, conservative, balanced and aggressive, as one line each with their trigger conditions, before any analysis. I own [number] shares of [TICKER], bought on [purchase date] at an average cost of [$cost], held in a [account type] account. My original reason for buying was: [thesis]. Look up the current price yourself. Beneath the verdict, show your work: the thesis tested against current fundamentals, recent news and valuation; the technical trend used as a risk gauge and not a verdict, skipped entirely with a note if this is a fund priced once daily; and the tax considerations I should raise with my advisor, including my holding period as of today and how lot selection would work. Then expand each exit path into full trigger conditions, specific and observable, so the decision gets made by rules rather than by mood.
09Risk Inventory & Tripwires
Lead with the three risks most likely to cost me money on [TICKER], each with its single sharpest tripwire — the observable indicator that would tell me it is becoming real. One line each, before anything else. This would be roughly [position %] of my portfolio over a [horizon] horizon. Beneath that, give me the full inventory: company-specific, industry, macro, valuation, competitive, regulatory and sentiment risks, plus the two most investors forget, customer concentration and key-person dependency. Pull concentration and dependency from the actual filings; if you can’t verify them, say so rather than estimating. Rank each low, medium or high, and justify the ranking against my position size rather than in the abstract. Give every risk in the inventory its own tripwire. Close with the one risk you think I am most likely to dismiss and shouldn’t.
10The Long-Term Thesis
Give me the long-term thesis for [TICKER] over [5 / 10 / 15] years in three sentences or fewer, then the kill criteria as a short list, before any supporting analysis. Beneath that, show what compounds and why you believe it: durable competitive advantage, leadership quality and their capital-allocation history, demand durability across cycles, pricing power, the innovation pipeline, financial strength, and credible expansion opportunities. Make every kill criterion something I could actually observe and put a date on, and say for each one roughly when I would expect to know. Close with the assumption in your thesis that is most likely to be wrong.
11Compare Three
Open with the ranking of [STOCK 1], [STOCK 2] and [STOCK 3] by risk-adjusted appeal, one line of reasoning each, and name which investor profile each one actually fits. Do that before any detail. My horizon is [horizon] and my risk tolerance is [risk]. Then give me the comparison table: quality of revenue growth (organic versus acquired), margins and returns on capital, debt load, valuation, moat durability, management quality, and the main risk to each. Beneath the table, expand on the two or three differences that actually drove the ranking rather than walking through every row. Where the three differ materially in size or in how much disclosure and coverage exists, say so, and state which metrics you had to substitute and why. Close with what would reorder the ranking.
Filled values highlight in brown; grey dashed chips still need a value. Every copy includes a short instruction block that tells the AI to use current data, name its sources, admit what it couldn’t verify, check whether your ticker is a company or a fund before it answers, and lead with the answer before the analysis rather than burying it at the bottom.
Scope. These prompts work on any listing. Name the market in the field above and the AI is told to use that market’s disclosure regime, filing names and reporting cadence rather than defaulting to United States rules. Tax is the exception: anything the prompts raise about holding periods, lot selection or account types is written to the United States federal system, and even there state treatment varies. Confirm your own jurisdiction with a local advisor before acting on any of it.
This tool changes. Version 2.0 added the instruction block, the fund check, and the vehicle prompt. If you want to hear when the next version ships, join the Outsider’s Letter. No obligation, and the tool stays free either way. See what changed →
Why this exists
The prompts are free. The framework is the book.
Every prompt on this page asks the questions an operator asks before committing capital: how the money is actually made, what the moat really is, what would kill the thesis. That discipline — entering industries you don't come from and out-executing the insiders — is the argument of An Outsider's Playbook, twenty short chapters, each closing with an operating principle you can use on Monday.
Disclaimer. This is an educational tool, not investment advice. Juan Vegarra is not a registered investment adviser, and nothing produced by these prompts — or by any AI assistant — is a recommendation to buy, sell, or hold any security. AI outputs can be wrong, outdated, or incomplete. Do your own diligence and consult a licensed financial professional before making investment decisions. Use of this tool is at your own risk.
The prompts evolve
Get the next version by email.
The toolkit stays free and ungated. Leave your email and you get each new prompt and refinement as it ships, plus one essay most weeks on operating and investing like an outsider.