How to Build an Investor Company One-Pager with AI
A source-controlled one-page company brief template, with a public-filing example and a workflow for separating refreshed facts from analyst judgment.
Published August 20, 2026 · Updated August 30, 2026
In this article
An investor company one-pager should let a portfolio manager understand the business, current financial shape, two or three operating drivers, valuation frame, thesis, risks, and next catalysts in one sitting. AI can assemble and refresh the factual layer. The covering analyst owns the thesis, scenario assumptions, risk ranking, and any recommendation. Every number needs a date, unit, period, and source that can be reopened.
This guide includes a public-source example built from Costco's fiscal 2025 Form 10-K and SEC company facts. It is a document-design example, not investment research or a recommendation. We build AllMind, which appears below as one workflow option, so read that section as our own case.
Use one page for orientation, not compression theater
A one-pager is not a miniature initiation report. Its purpose is to make the next conversation better. If it contains fifteen charts, a paragraph of market history, and a six-bullet thesis, it has stopped doing that job.
Use this page contract:
| Block | Maximum space | Required content | Owner |
|---|---|---|---|
| Identity and as-of line | 2 lines | Name, ticker, reporting currency, fiscal year-end, price timestamp | Automated, then checked |
| Business model | 3 sentences | Who pays, what they buy, and the economic engine | AI draft; analyst edits |
| Financial spine | 5 to 7 rows | Revenue, profit measure, cash flow, leverage, and company-specific KPIs | Automated from approved data |
| What drives the next 12–24 months | 3 bullets | Operating variables, with current evidence | Analyst |
| Valuation frame | 3 rows | Chosen multiples or DCF inputs, date, peer/history context | Analyst chooses method |
| Thesis and disconfirming evidence | 2 + 2 bullets | Why the view differs and what would prove it wrong | Analyst |
| Catalysts and watch dates | 3 rows | Event, date/window, evidence expected | Shared |
| Sources | 3 to 6 links | Filing, release, model/data timestamp, key transcript | Automated |
The strict space limit forces a distinction between facts needed for orientation and facts that belong in the underlying research file.
A copyable one-pager schema
The following schema works in Word, PowerPoint, a research management system, or a generated report.
Header
Company | Ticker | Currency | Price as of [timestamp] | FY end | Analyst | Last factual refresh
Business in three lines
- The customer and product or service.
- The revenue and profit mechanism.
- The operational constraint or advantage that matters most.
Financial spine
| Metric | FY-2 | FY-1 | LTM / latest FY | Next-year estimate | Source status |
|---|---|---|---|---|---|
| Revenue | Filed / estimated | ||||
| Operating profit or EBITDA | Filed / adjusted | ||||
| Net income | Filed | ||||
| Operating cash flow | Filed | ||||
| Capital expenditures | Filed | ||||
| Sector KPI 1 | Defined by company | ||||
| Sector KPI 2 | Defined by company |
Do not place historical actuals and consensus estimates in the same visual treatment. An estimate should never look filed merely because it shares a row.
Decision blocks
- Variant view: one sentence stating what the market appears to expect and what the analyst expects differently.
- Evidence for: two observations tied to sources.
- Evidence against: two observations that weaken the thesis.
- Next test: the next disclosure or operating datapoint that will update the view.
- Valuation: method, date, key input, and sensitivity. Do not show a target without the assumptions that create it.
Worked example: the factual layer for Costco
We pulled Costco's fiscal 2025 standardized facts from the SEC's company-facts JSON on August 30, 2026 and checked them against the company's fiscal 2025 Form 10-K, filed October 8, 2025.
| Factual one-pager field | Fiscal 2025 value | Evidence note |
|---|---|---|
| Revenue | $275.235 billion | SEC standardized fact, fiscal year ended August 31, 2025 |
| Operating income | $10.383 billion | SEC standardized fact |
| Net income | $8.099 billion | SEC standardized fact |
| Operating cash flow | $13.335 billion | SEC standardized fact |
| Capital expenditures | $5.498 billion | SEC standardized cash-flow fact |
| Warehouses | 914 | Company disclosure in Business section |
| Membership fee revenue | $5.323 billion | Company table; reported in USD millions |
| Renewal rate | 92.3% US and Canada; 89.8% worldwide | Company-defined trailing calculation |
Those rows are suitable for automated assembly because the filing states the values and definitions. They still require care. Costco says the renewal-rate calculation captures renewals during a specified trailing period and excludes affiliates of business members. A one-pager that reduces the row to “90% retention” has lost both precision and definition.
The business summary can be drafted from the same filing: Costco operates membership warehouses and e-commerce sites, uses a limited product selection and high inventory turnover, and earns membership fees alongside merchandise sales. The analyst should edit the language because deciding which part of the model is most important is already analysis.
This example stops before valuation and thesis. Price, estimates, peer multiples, and a recommendation need dated market data and analyst assumptions that the public filing does not supply.
Give the AI a source hierarchy
Without an explicit hierarchy, a system may use an old investor presentation when a newer filing exists or blend a third-party estimate into a historical row. A sensible order is:
- filed statements and notes for historical financials;
- filed or company-published releases for the most recent event;
- company-defined KPI tables and reconciliations;
- entitled consensus or the analyst's model for estimates;
- call transcripts for explanation and management language;
- reputable industry sources for external operating data;
- internal research for the thesis, prior questions, and house definitions.
The SEC explains how its submissions and company-facts APIs differ. That distinction should survive in the source record. A structured fact is excellent for a financial row, while the filing passage is often necessary for a KPI definition or accounting change.
Refresh facts and judgment on different clocks
Not every block needs to change after every event.
| Event | Refresh automatically | Require analyst review |
|---|---|---|
| Earnings release or filing | Financial spine, as-of line, disclosed KPIs, catalyst status | Thesis, scenario assumptions, valuation inputs |
| Material 8-K | Event line and affected fact | Whether the thesis or risk ranking changes |
| Price move | Price and market-value fields | Whether valuation or catalyst interpretation changes |
| Peer report | Peer comparison evidence | Read-through to the company |
| Annual deep review | Full factual layer and source audit | Rewrite business summary, thesis, risks, and monitoring plan |
Display two timestamps when needed: “facts refreshed” and “analyst view reviewed.” A current share price does not make a six-month-old thesis current.
Product requirements for a one-pager workflow
The document can be generated with a general assistant if an analyst supplies a clean source pack and verifies every field. That may be enough for an occasional page. A coverage-wide process has stricter requirements:
- reusable field definitions and house formatting;
- access to the approved filing, fundamentals, estimates, and internal notes;
- period and entity mapping across the coverage list;
- source pointers preserved in the output;
- scheduled or event-based refresh;
- a diff showing which facts and sentences changed;
- approval before the page is distributed.
We built AllMind for that connected workflow, and it can generate documents in a supplied template from external and internal research. The filings, fundamentals, and estimates the page needs are already licensed inside the platform, so only the firm's own notes and models require a connection, and onboarding is sales-led for that step. A market-data terminal or low-cost fundamentals service may be sufficient when the goal is a manually written snapshot on a few companies.
Quality checks before distribution
Run these checks on every refreshed page:
- exactly one as-of timestamp for price and market data;
- historical, estimated, and calculated values labeled differently;
- fiscal periods used consistently;
- company KPI definitions linked and unchanged, or the change explained;
- no valuation output without its method and key assumption;
- every catalyst has a date or observable condition;
- thesis and risk text carries the analyst's review date;
- every displayed number opens to a filing, dataset row, or model cell.
We did not test document generators or compare their output quality. The Costco values above are a public-source extraction, not a live product run, and later amendments could change the SEC data. The template does not tell a reader what to own. It creates a compact, inspectable record of what the analyst currently knows and believes.
Sources and methodology
- Costco fiscal 2025 Form 10-K, for business description, warehouse count, membership fees, and renewal-rate definition.
- Costco SEC company facts, for standardized financial facts captured August 30, 2026.
- SEC EDGAR API documentation, for source mechanics and API scope.