AI Output Verification Prompt Companion
Systematically verify AI-generated content before publishing, presenting, or making decisions.
Markdown + PDF · v1.0.0 · Model-agnostic
What’s Inside
5 Verification Families
Organized by professional process, not alphabetically.
EXTRACT
4 promptsPull claims out of AI content, classify them, flag fabrication patterns
- EX-01 — Claim Extraction
- EX-02 — Suspicious Claim Detection
- EX-03 — Source Inventory
- EX-04 — Claim Classification Matrix
VERIFY
3 promptsCheck specific claims against evidence and assess source quality
- VR-01 — Single-Claim Verification
- VR-02 — Statistical Verification
- VR-03 — Source Hierarchy Assessment
ASSESS
3 promptsScore overall document reliability and calculate confidence
- AS-01 — Confidence Scoring
- AS-02 — Document Reliability Assessment
- AS-03 — Temporal Accuracy Check
CHALLENGE
3 promptsStress-test your verification — find what you missed
- CH-01 — Adversarial Review
- CH-02 — Citation Trap Scanner
- CH-03 — Publication Blocker
REPORT
2 promptsSynthesize findings into professional summaries and recommendations
- RP-01 — Verification Summary
- RP-02 — Publish/Revise/Rewrite Recommendation
Sample: Suspicious Claim Detection (EX-02)
One of 15 prompts in the library. This scans AI content for 6 categories of fabrication.
Review the following AI-generated content for claims that
match known AI fabrication patterns. Flag anything that matches
these categories:
FABRICATION CATEGORY 1 — STATISTICAL INVENTION
- Round numbers presented with false precision ("exactly 47%")
- Statistics without a named, dated source
- Numbers that are suspiciously convenient for the argument
FABRICATION CATEGORY 2 — SOURCE FABRICATION
- Studies, papers, or reports that sound real but may not exist
- Authors with plausible-sounding names attached to unverifiable work
- Organizations that may be invented
FABRICATION CATEGORY 3 — QUOTE FABRICATION
- Direct quotes that are too perfectly articulate
- Quotes that align too neatly with the article's thesis
FABRICATION CATEGORY 4 — TEMPORAL DISPLACEMENT
- Facts that were true at one time but may be outdated
- Statistics from a specific year applied to the current year
FABRICATION CATEGORY 5 — CAUSAL INFLATION
- Correlation presented as causation
- Single-study findings presented as established consensus
FABRICATION CATEGORY 6 — CONFABULATED CONTEXT
- Plausible backstories added to real events
- Details about processes that sound right but may be invented
Content:
[PASTE CONTENT]
For each flagged item:
1. Quote the suspicious text
2. Identify the fabrication category (1-6)
3. Explain specifically WHY it is suspicious
4. Rate the likelihood: Possible / Probable / Very likely
5. Suggest how to verify it
If nothing looks suspicious, say so — do not fabricate problems.Who This Is For
AI Compatibility
Recommended: Use a different model than the one that generated the content. Web access is optional but useful for source verification prompts.
License Options
Personal
$29
1 user
Individual use within your own business
Professional
$39
1 user
Individual commercial use including client work
Team
$59
Up to 10 users
Shared use within one organization
Consultant
$79
1 consultant
Repeated use across multiple client engagements
All licenses include modification rights for your own authorized use. You may customize prompts to your domain. Redistribution and resale are not permitted.
Frequently Asked Questions
What AI do I need?
Any capable AI model — ChatGPT (GPT-4o+), Claude (3.5+), Gemini (Pro+). We recommend using a different model than the one that generated the content you're verifying. Web access capability is optional but useful for source verification.
How is this different from free verification prompts?
Every prompt embeds specific methodology: 6 fabrication categories, claim taxonomy, source hierarchies, and a weighted confidence scoring formula. We benchmark every prompt against "just ask AI" — if a simple request gets you 80% of the way, we make it free instead.
How many prompts are included?
15 prompts organized into 5 families (Extract, Verify, Assess, Challenge, Report). Plus 2 worked examples, a troubleshooting guide, and a quick reference card.
Can I modify the prompts?
Yes — you're encouraged to adapt them to your domain. The START HERE guide explains which sections are safe to modify and which encode specific methodology.
Do I have to use all 15 prompts every time?
No. For a quick check, use EX-02 alone. For standard verification, use EX-02 → VR-01 → CH-01. The full 5-family process is for thorough verification of high-stakes content.
What format do I get?
Markdown files (easy to copy) and a PDF (professional reference). Each prompt is formatted as a Prompt Card with name, job, inputs, prompt text, expected output, and quality check.
Does KaryoWorks run the AI for me?
No. You use your own AI account. KaryoWorks provides the professional instruction architecture — you provide the execution environment. No API costs from KaryoWorks.
What's the relationship to the AI Research Verification System?
The Verification System ($99) is the complete methodology — scoring spreadsheets, evidence matrices, report templates, and implementation guides. This Prompt Companion ($29) helps you execute parts of that methodology using AI. They work independently but are stronger together.
What if AI models get so good they don't hallucinate?
That would be great — and would reduce the need for these prompts. But verification demand is counter-cyclical: the more AI output in the world, the more verification matters. The methodology (what to check, how to score, how to assess sources) survives regardless of model quality.
Goes deeper with the complete system
This prompt companion helps you execute verification with AI. The full system provides the complete methodology.
AI Research Verification System
Scoring spreadsheets, evidence matrices, verification report templates, and implementation guides.
$99
DecisionsAI Decision System for Business
When verification supports a business decision — the complete decision framework.
$149
EverythingComplete AI Professional System
All three systems in one bundle at 24% off. Includes this prompt companion.
$249