Collect
Store market, filing, news, and industry inputs with their source and snapshot date.
Ask about a company, constellation, market signal, or strategic choice. The product shows its sources, runs repeatable valuation or risk calculations, marks missing fields, and saves the result as a report for you to review.
The five-step example uses fixed data. Each button opens the product screen used for that step.
Choose a company, industry, comparison, or decision. The Workbench opens the matching research view.
Open Ask LaurenceYangWhat supports this satellite operator’s moat, and what could weaken it?
Understand the companyBusiness, financials, and valuation
Check the riskEvidence, events, and counterpoints
Review the numbers, sources, assumptions, and missing-data notes before trusting the conclusion.
Open company researchPut companies side by side without changing the evidence rules between them.
Open comparison workflowA live comparison needs provider coverage for at least two companies.
The first screen stays focused on the question; the extra analysis remains separate.
Open advanced analysisChange the model or screen a wider set of companies.
Turn the research into a report you can review, update, and export to PDF.
The other pages handle satellite events, dated briefs, company topics, and concept research.
Ask a question, compare companies, run a valuation, test a thesis, and save a report in one workspace.
Classify events, test thesis impact, and compare CEO and investor views.
Read the current evidence snapshot, changes, catalysts, and source freshness.
Track the thesis, counter-evidence, risks, and next review actions.
Focused topic pages for versioned market snapshots, trends, alerts, peers, moat, and bull/bear views.
Browse concepts, entities, relationships, and source references.
Data tools produce the numbers, and the language layer summarizes their output. This public site uses dated snapshots and fixed, labeled examples; it has no live provider or model connection.
Store market, filing, news, and industry inputs with their source and snapshot date.
Normalize identifiers, cross-check sources, calculate DCF, peers, screens, and risk signals.
Summarize the tool results in plain English. If no AI model is connected, show a fixed sample response instead.
Show assumptions, source dates, missing data, and counter-evidence for human review.
Sources, calculations, missing-data labels, and reports follow one set of controls.
The deployment contains fixed, labeled examples. It contains no live provider, model API, credential, or restricted Wiki full text.
The personal, non-commercial release contains dated examples and public previews. Provider adapters, local APIs, credentials, and restricted Wiki full text are not deployed.
Open the Workbench to follow the product flow. How It Works explains the calculations, data controls, and my contribution.