Skip to main content Skip to main content
Not investment advice. Educational reading. See Disclaimer.
L.8 · INTERMEDIATE · 5 MIN

Research with an AI Assistant: Source, Date, and Refusal

AI assistants can now look things up for you: pull a company's filings, summarize an annual report, line up what several lenders say about one loan. For a patient researcher that is a real help. It also opens a new way to be wrong, because an assistant can write a fluent, confident sentence with no document behind it. The answer is not to avoid assistants. It is to apply three habits careful readers of filings already use to every answer you get: check the **source** (which filing), check the **date** (what moment a number describes, and when it became public), and respect a **refusal** (an honest 'I can't answer that' beats a filled-in guess). The examples below are real responses from Oxford Ledge's own research tools, each labeled with the day it was captured, 2026-09-27. They are dated records, not live data, and the numbers in them will have moved by the time you read this. The habits work the same way with any assistant, data site, newsletter, or tip from a friend.

Quiz · 5 questions ↓

Three habits and the question behind each

HabitThe question to askWarning sign
Check the sourceWhich filing is this from: which company, which form (10-K, 10-Q, 20-F), and is it the filing's own number or someone's paraphrase?A figure with no document behind it, or 'according to reports'
Check the dateWhat date does the number describe, and when was it filed? Do all the rows share the same dates?One 'current' figure assembled from documents filed weeks apart, with no dates shown
Respect a refusalDid the answer say it could not find or compute something, and did it say why?Asking the assistant to 'just estimate it', or reading an empty result as a zero

Habit one: ask which filing

Captured 2026-09-27. Asked for a business summary of Costco (COST), the tool returned a paragraph plus labels saying where it came from: source_form 10-K, source_period 2025-08-31, and a note that the paragraph is a 'model-written paraphrase of the filing, not filing text'. Those labels answer the source question in full. The summary came from Costco's annual report for the fiscal year that ended August 31, 2025, and because it is a paraphrase, anything you want to quote as Costco's own words has to come from the 10-K itself. Notice the date too: on the capture day, that annual report described a year that had closed more than twelve months earlier, and it was still Costco's newest 10-K, because the report for the year that had just ended had not yet been filed. When an assistant gives you a figure without labels like these, ask for them: which company, which form, which period, and a link to the filing. If it cannot name a source, treat the figure as unverified, however precise it looks.

Habit two: every number carries two dates

Captured 2026-09-27. Asked how the lenders to Coupa, a business-software company whose loans are held by private-credit funds, value those loans, the tool returned 8 business development companies (BDCs: funds that lend to private companies and report a fair value for each loan every quarter). Each lender's mark, in cents per dollar of the loan's face value (averaged by size when a lender holds more than one piece of the debt), ranged from 94.85 at Ares Capital (ARCC, filed 2026-07-29) to 100.0 at Carlyle Secured Lending (CGBD) and Morgan Stanley Direct Lending Fund (MSDL), both filed 2026-08-06; the other five fell within that range. Every one of those marks is as of the same quarter-end, June 30, 2026. The filings behind them reached the public on different days, from July 29 to August 7, and the tool said so in its own summary: the marks 'span DIFFERENT filing dates (2026-07-29 to 2026-08-07)' and are 'NOT necessarily contemporaneous'. That warning is the safe default, because filings weeks apart can describe different quarters; checking the as-of dates is what tells you these rows line up. Each number carries two dates, the moment it describes and the day it became public. A table that mixes filing dates is still useful, as long as you read it that way.

Reading rows that span filing dates

An assistant reports that five lenders value the same private loan at between 96 and 100 cents on the dollar. Every value is as of June 30. The lowest, 96, comes from the lender that filed first, on July 29; the others filed through August 7. Which reading do the dates support?

Habit three: a refusal is an answer

Captured 2026-09-27. Asked for intrinsic-value estimates for Taiwan Semiconductor (TSM), the tool returned no numbers at all. Each of its three models came back empty with the reason no_us_gaap_annual_facts, and the summary explained why: companies that report under IFRS rather than US accounting rules 'are the usual cause -- the financials exist; these tools read us-gaap concepts only.' That is a good answer. It tells you where the tool's reach ends and that the numbers exist elsewhere. For a foreign company like TSM, that means its annual report on Form 20-F, which you can read directly on EDGAR. The weak response is to push back with 'just estimate it.' An assistant asked to fill a gap can produce a figure that looks exactly like a real one, and a number with no filing behind it is the thing habit one exists to catch. Treat a clear refusal with a reason as a sign of care, and follow its pointer instead of overriding it.

Habit three, continued: an empty answer is not a zero

Captured 2026-09-27. The same lender tool, asked about 'costco', returned zero rows and one sentence: 'not in the corpus under this borrower_norm as a funded debt position (equity/unfunded only, or a different key).' Read what it claims and what it does not. It says the private-credit dataset holds no funded Costco loan under that name. It does not say Costco has no debt; Costco's own 10-Q lists billions of dollars of long-term debt. The mistake to avoid is turning 'found nothing here' into 'there is nothing', whether you make it or your assistant does. When an answer comes back empty, ask two questions: what exactly was searched, and under what name?

Audit one figure you have already seen

Pick one figure you have seen quoted recently: in an article, a post, a stock page, or an assistant's reply. Answer three questions about it. Source: which filing, from which company, and is the figure the filing's own number or someone's restatement of it? Date: what date does it describe, and when was that filing made public? Refusal: did anyone along the way fill a gap with an estimate, or read a missing number as zero? If you cannot answer the source question after ten minutes on SEC EDGAR, write the figure down as unverified rather than as fact.

Source, date, refusal

So far

An assistant is only as trustworthy as the checks you run on its answers. Ask which filing a number came from, and whether it is the filing's own figure or a paraphrase. Ask what date it describes and when it was filed, and notice when rows mix filing dates. When an answer says it cannot compute something, or comes back empty, take it at its word and follow its pointer rather than asking for a guess. These are the same habits that make you a careful reader of filings; an assistant puts them to work more often.

Optional: practice in your own assistant

If you would like to practice these habits with the tools quoted above, you can connect any assistant that supports the Model Context Protocol (MCP), an open standard for letting an assistant call outside tools, by following the guide at oxfordledge.com/mcp.

Check your understanding

Sit with the ideas.

You ask an assistant whether any private-credit funds lend to a large, well-known retailer. It replies that it found no rows under that name in its lender dataset, which covers funded loans only, and that the retailer may be listed under a different name. Which reading does that answer support?

Why:
Continue this lesson in the app →See it on a real ticker →