Trust, citations, and auditability

Source-Checked Stock Research

Use ThesisCheck for stock research where every material claim is tied to dated evidence and missing proof is shown plainly.

Reviewed 2026-06-27Updated 2026-06-27
The question

I want stock research with citations I can inspect instead of confident AI summaries that do not show where the claims came from.

What this answers

Source-checked stock research ties material claims to dated evidence, separates exact quotes from summaries, and marks unsupported proof gaps. ThesisCheck applies that structure to a user-provided stock thesis.

Definition

What source-checked stock research means

Source-checked stock research is research where the important claims can be traced back to dated public evidence. The useful unit is not just the paragraph; it is the claim, source, date, locator, and evidence status.

  • Exact source language is labeled as a quote.
  • ThesisCheck-authored interpretation is labeled as a summary.
  • Insufficient evidence is labeled as a gap, not promoted into certainty.
Source quality

Primary sources, secondary sources, and unsupported claims

The point is not to add citations for decoration. The point is to show whether a claim can be checked from public evidence.

01
Prefer primary evidence

Filings, annual reports, and investor materials matter because they are tied to the company record.

02
Keep summaries separate

A summary can be useful, but it should not be disguised as source language.

03
Mark unsupported evidence

When the public record shown does not prove a claim, the page should say so plainly.

Mini example

Example source-ledger snippet

Input thesis

Microsoft's AI infrastructure spending is justified by cloud revenue growth.

Checked claim

The public demo can verify cloud growth from the annual report, but not a standalone AI infrastructure ROI calculation.

Mixed2025-07-30

Evidence summary: The annual report supports Azure and other cloud services growth, while the abridged public evidence shown here does not isolate product-level ROI for AI infrastructure investments.

Microsoft FY2025 Annual Report · 2025-07-30 · Segment Results and abridged demo corpus
What the user learns

A source-led review should show both sides of the evidence boundary: what the source supports, what the summary infers, and what remains unsupported.

Comparison

Generic AI summary vs source-checked thesis review

WorkflowGeneric AI summarySource-checked thesis review
Evidence trail

May provide a fluent answer without showing where each claim came from.

Shows source label, date, locator, and whether the evidence is quote, summary, or gap.

User problem

Answers a broad prompt about a company or market.

Tests a specific thesis against public company evidence.

Failure mode

Can sound certain while blending support, context, and speculation.

Separates support, contradiction, and missing evidence.

Non-advice boundary. ThesisCheck provides descriptive, source-checked company research only. It does not provide buy, sell, hold, rating, sizing, or price-target recommendations.