Basic
$199/moBelow DA/DR targetNot a fitExcluded from recommendation
AGENCY AI TRANSFORMATION
SITETRAIL CERTIFIED
Your learning path: 16 lessons.
Module 03 / Lesson 11 · Read, apply, check
Run a reproducible check that distinguishes supported facts from assumptions, estimates and unsupported claims.
List each consequential statement, its type, source and verification status. Include numbers, comparisons, dates, product capabilities, customer outcomes and quotations. Do not limit the register to sentences that look technical. “The first provider” and “all customers” may be harder to substantiate than a straightforward revenue figure. Mark supported, needs clarification, contradicted or remove. AI can help extract candidate claims, but the editor decides which assertions need review and checks the source itself rather than accepting the tool’s confidence as evidence.
Check units, denominators, periods and rounding. A change from 20% to 30% is ten percentage points and a 50% relative increase; the two descriptions answer different questions. Confirm whether a figure is an average, median, total or forecast. Reproduce a sample calculation independently and inspect edge cases such as zero denominators and missing values. Keep the calculation note with the source ID. A beautiful chart can preserve a wrong denominator, so verify the underlying measure before approving the visual or headline.
Open the original document or recording and locate the relevant passage. Check publication date, author, definitions and whether the source reports direct evidence or repeats someone else’s claim. Five articles repeating one press release are not five independent confirmations. A source may support a narrower statement than the draft makes. For quotes, compare the exact wording and context; for product claims, check current authorized documentation and the relevant version. Record unresolved discrepancies as questions instead of choosing whichever source best supports the desired story.
Resolve or remove material issues before final approval, and record who checked what. Keep a short list of limitations that must survive editing and adaptation. Review the headline, image captions, chart labels and promotional excerpts for stronger wording than the body supports. If a defect is discovered after release, identify affected versions and destinations, assess the correction needed and assign its owner. Verification is not a ceremonial final read; it is a documented process that makes both prevention and correction possible.
List the claim, its type and the consequence if wrong.
Find the original source and exact supporting passage.
Check calculation, context, scope and quote fidelity.
Approve, qualify, return or remove; preserve the decision record.
A fictional client report shows trial conversion rising from 20% to 30%. The AI draft calls this a ten percent improvement, then later says conversion doubled. The editor checks the underlying counts, confirms the same conversion definition and period length, and calculates a ten-percentage-point rise or 50% relative increase. The claim register records both valid descriptions and the reason for choosing percentage points in the article. A chart is corrected to use the same measure. The editor also discovers that three supporting news articles repeat the client’s own announcement, so the draft no longer describes them as independent validation. The final story keeps the observed result but removes the false precision and exaggerated corroboration introduced during drafting.
Another team asks a second chatbot whether the first chatbot’s article is accurate. The second tool repeats the same plausible number and supplies a link that mentions the company but not the claimed result. The team treats agreement as verification. When the client questions the calculation, nobody can reproduce it. Model agreement has replaced inspection of the original evidence, leaving the same unsupported assertion dressed in an additional layer of confidence.
Fact-checking is not simply finding a web page that contains similar words. The source must support the exact claim, period and scope. Independent recalculation and direct inspection are stronger than asking for reassurance. Equally, a source’s prestige does not make every inference drawn from it valid; separate what it reports from what the agency concludes.
Some claims require specialist review or cannot be established from the available material. Escalate those instead of treating a general-purpose AI answer as expert validation. Keep forecasts and estimates clearly labelled, and do not present an illustrative teaching example as a measured client result or a verified market fact.
Do this today · 30 minutes
Check the conversion example and create a five-claim register.
A document or spreadsheet and the illustrative case in this lesson. Use synthetic or already approved material for practice; no paid account or real client data is required.
Before: List the claim, its type and the consequence if wrong.
After: Approve, qualify, return or remove; preserve the decision record.
Notes stay in this browser. Download a copy to keep them.
Test this worksheet on one bounded assignment. Record the reviewer’s findings, revise the method and assign the next review date.
Before you move on
A client’s conversion rate rises from 20% to 30% using the same definition and comparable periods. The draft says it doubled, and two AI tools agree. What should the editor record after checking the underlying counts?
Next lesson →Associated Press: newsroom standards for AI
PRSA: professional ethics resources
Lesson edition 1.0 · References reviewed 16 October 2026 · Practical exercise reviewed 16 October 2026. Examples are illustrative teaching scenarios.
Sitetrail service · NewsPass
Put your SEO, PR and generative engine optimization strategy into practice. NewsPass supports search visibility, media presence, brand credibility and discoverability in AI answers.
Compare the delivery costs for your campaign using the same calculator as the Sitetrail homepage. Choose a target authority range, your typical placement cost and expected monthly volume.
Illustrative cost comparison using October 2026 plan prices; check current terms before purchase. Assumes one eligible placement per credit and comparable selected opportunities, not a prediction of rankings, revenue, coverage or AI citations. Actual suitability depends on audience fit, editorial requirements and the current plan terms. DA/DR are third-party metrics, not Google scores.
Choose publications available on your plan, prepare useful content for the audience and submit it for editorial review. Track accepted placements in your dashboard and evaluate them against your campaign goals.
Original evidence. Human editorial judgment.
Use the worksheets to turn each lesson into a real decision. Improve the method as you learn from client evidence, editorial checks and delivery results.
Course outline · Free assessment · Privacy
No account is required. Progress, first-attempt lesson answers, worksheets and assessment results are stored in this browser when storage is available. They do not sync across devices. Download copies to retain them; clearing browser data removes local records.
Your free trial starts automatically. No license key is needed yet.