Find AI claim substantiation tools and buyer checklists
Use these guides to review public AI product claims before a buying decision. Pick the closest checklist, review the evidence gaps and buyer questions, then check the exact vendor wording in the product claims checker. Start with Quick reference or Choose by task when titles look similar; use narrower pages after you know the claim pattern.
Use this index
The guides help you classify a claim before checking exact wording.
-
Start with the wording pattern
Look for the part of the vendor wording that creates the evidence burden: a percentage, automation promise, compliance reference, ROI number, superlative, or vague AI label.
"98% accurate" maps to accuracy evidence; "fully automated" maps to human-review boundaries.
-
Open the closest guide
Use the guide to see source-backed evidence gaps, buyer questions, and safer wording boundaries for that claim type.
Guides explain what to ask for before you rely on the wording.
-
Check the exact claim
After you identify the pattern, paste the exact vendor wording into the checker to generate a Claim Receipt for that specific claim.
The checker is the action step; the guide is the pattern map.
Choose by task
Use this when adjacent checklist pages look similar. Pick the guide by the review job in front of you.
-
One claim sentence
Open substantiation guide →Use the substantiation guide.
Map one accuracy, compliance, automation, ROI, benchmark, or superlative claim to the evidence a buyer should request.
-
One vendor page or claim set
Open vendor checklist →Use the vendor evidence checklist.
Walk a product page sentence by sentence before turning the exact wording into a receipt.
-
Pre-purchase comparison
Open due diligence checklist →Use the due diligence checklist.
Compare claims across sales decks, demos, questionnaires, DDQ answers, and public product pages before a buying decision.
-
High-burden marketing phrase
Open marketing checklist →Use the marketing claims checklist.
Spot superlatives, vague AI-powered labels, outcome guarantees, and compliance references before checking the exact phrase.
Quick reference
Summary pages covering multiple claim types — use these first to find the right pattern.
-
AI claim risk phrases
Nine high-risk AI claim patterns — accuracy, automation, compliance, ROI, uniqueness — with evidence needed and official source basis for each.
Read guide →Pattern / review use
All claim types · Evidence needed · Buyer questions
"Fully automated" · "98% accurate" · "Bias-free AI" · "Guaranteed ROI"
Regulatory sources
Catalog source context for AI claim evidence burden.
-
FTC AI claims
How FTC substantiation standards apply to AI accuracy, outcome, and automation claims in marketing copy.
Read guide →Pattern / review use
Accuracy / Performance · ROI / Outcome · Automation / Replacement
"sue for assault without a lawyer"
-
SEC AI washing
AI use and performance wording from SEC source examples. Use as source context, not investment advice.
Read guide →Pattern / review use
AI usage claims · Performance / Returns · First / Only / Best
"make our artificial intelligence smarter"
-
EU AI Act transparency claims
Transparency, AI disclosure, and handoff wording under EU AI Act source context, with buyer questions for public AI product claims.
Read guide →Pattern / review use
Compliance / Safety · AI disclosure · Human handoff
"Fin AI Agent is more than a chatbot"
-
FTC Operation AI Comply
Named Operation AI Comply sweep patterns from AI lawyer, review-generation, ecommerce income, and detector cases.
Read guide →Pattern / review use
FTC sources · Claim pattern · Evidence records
"the world's first robot lawyer"
Evidence and performance claims
Accuracy, benchmark, outcome, and evidence-support wording — what a buyer should ask for.
-
AI claim substantiation
Map AI accuracy, compliance, bias, and automation wording to the evidence a buyer should request.
Read guide →Pattern / review use
Evidence needed · Claim type · Source date
"system can't be tricked by a photo or video image"
-
AI vendor claim evidence checklist
Step-by-step evidence checklist for a single vendor page or pasted claim set—not a full procurement review.
Read guide →Pattern / review use
Evidence checklist · Buyer questions · Scope limits
"use artificial intelligence to boost earnings for consumers' e-commerce storefronts"
-
AI accuracy claims
Accuracy and performance wording checked against field evidence, benchmark scope, subgroup results, and testing limits.
Read guide →Pattern / review use
Accuracy / Performance · Field evidence · Benchmark scope
"one of the highest accuracy rates on the market"
-
AI detector accuracy claims
Numeric accuracy, benchmark scope, false positive / false negative rates for AI detection tools.
Read guide →Pattern / review use
Accuracy / Performance
"98 percent accurate"
-
AI customer support agent accuracy claims
Resolution-rate, deflection, hallucination, handoff, and accuracy wording on support-agent pages, with buyer evidence questions.
Read guide →Pattern / review use
Accuracy / Performance · Resolution / deflection · Escalation
"resolve 80%+ of customer and employee interactions instantly across any channel"
-
Bias-free AI claims
Bias-free and fairness wording, with questions about tested groups, error rates, monitoring, and limits.
Read guide →Pattern / review use
Accuracy / Performance · Fairness evidence · Subgroup results
"free of gender and racial bias"
-
AI ROI and cost-saving claims
AI revenue, savings, productivity, and outcome wording checked against customer result distribution and cost basis.
Read guide →Pattern / review use
ROI / Outcome · Cost basis · Customer results
"quickly earn thousands of dollars a month in passive income"
-
AI chatbot and LLM accuracy claims
When an AI assistant or LLM claims professional-level output, this guide shows what task scope, qualified comparison, and failure-condition records to ask for.
Read guide →Pattern / review use
Task scope tested · Qualified comparison evidence · Failure conditions disclosed · Human review boundary
"world's first robot lawyer — generate perfectly valid legal documents in no time"
-
First only best AI claims
Uniqueness and superlative AI claims checked against market scope, comparison set, and point-in-time evidence.
Read guide →Pattern / review use
First / Only / Best · Market scope · Comparison evidence
"first regulated AI financial advisor"
-
Hallucination-free AI claims
No-hallucination, low-hallucination, source-grounded, and answer-validation wording mapped to benchmark and monitoring evidence.
Read guide →Pattern / review use
Hallucination rate · Grounding · Error monitoring
"very low hallucination rate (<1%)"
-
AI benchmark claims
AI benchmark scores, state-of-the-art wording, leaderboard claims, and production-performance transfer claims mapped to buyer questions.
Read guide →Pattern / review use
Benchmark method · Evaluation scope · Production fit
"scoring 74.9% on SWE-bench Verified and 88% on Aider polyglot"
Automation and wording claims
AI-powered, replacement, and fully automated wording — where scope and human review matter.
-
AI-powered meaning
Vague AI-powered, AI-native, built-in AI, and agentic AI descriptors mapped to workflow evidence and buyer questions.
Read guide →Pattern / review use
Vague AI-powered · Workflow role · Buyer questions
"Power your business with built-in AI marketing tools"
-
AI replacement claims
Professional-replacement and no-human-needed AI claims, with questions about task scope, testing, and review boundaries.
Read guide →Pattern / review use
Automation / Replacement · Human review · Task scope
"the world's first robot lawyer"
-
Fully automated AI claims
Fully automated, no-human-review, and AI agent write-action wording, with questions about approval workflows, rollback, audit logs, and review limits.
Read guide →Pattern / review use
AI agent actions · Approval workflow · Rollback logs
"eliminated the need for human order-taking"
-
AI washing examples
Common vague and overstated AI claim patterns with evidence gaps, evidence signals, and wording boundaries.
Read guide →Pattern / review use
Vague AI-powered · First / Only / Best · Automation / Replacement
"AI content detector can identify AI-generated text"
-
Human review boundary claims
Expert-reviewed, human-in-the-loop, handoff, and oversight wording—what evidence buyers should request when a claim implies human involvement without defining it.
Read guide →Pattern / review use
Automation / Replacement · Human handoff · Review boundary
"Fin will provide instant responses and a robust handover experience"
Security, privacy, and compliance claims
Privacy, security, and compliance wording — what evidence can support the public claim.
-
HIPAA-compliant AI claims
Healthcare AI privacy and security wording, with buyer questions about vendor role, data flow, safeguards, and scope limits.
Read guide →Pattern / review use
Compliance / Safety · Vendor role · e-PHI scope
"HIPAA compliant"
-
Secure AI assistant claims
Security and private-by-design AI assistant wording, with questions about data retention, admin controls, model access, and safeguards.
Read guide →Pattern / review use
Compliance / Safety · Security controls · Data use
"Enterprise-grade privacy and security controls"
-
AI cybersecurity claims
AI-powered threat detection, vulnerability discovery, incident response, SOC automation, and frontier-model cybersecurity wording mapped to evidence questions.
Read guide →Pattern / review use
AI cybersecurity · Threat detection · Vulnerability evidence
"generative AI-powered security solution"
-
AI transparency claims
Transparent, explainable, auditable, traceable, disclosure, and source-visibility wording mapped to buyer evidence questions.
Read guide →Pattern / review use
Compliance / Safety · Vague AI-powered · Disclosure boundary · Traceability evidence
"Fin AI Agent is more than a chatbot"
-
AI hiring and screening tool claims
When an AI hiring tool claims fair, objective, or consistent evaluation, this guide shows what bias audit records, candidate disclosure, and FCRA evidence to ask for.
Read guide →Pattern / review use
Bias audit record · Candidate disclosure · FCRA applicability · Audit publication
"ensuring consistent, fair evaluation for every candidate"
-
AI subprocessor and model-provider claims
Subprocessor lists, underlying model providers, prompt-routing paths, support access, and regional-processing exceptions mapped to buyer questions.
Read guide →Pattern / review use
Subprocessors · Model provider path · Data routing
"between OpenAI and its service providers"
-
AI training data claims
Training-data source, provenance, cutoff, customer-data use, and continuous-learning wording mapped to buyer evidence questions.
Read guide →Pattern / review use
Training data · Data provenance · Model updates
"We don't train our models on your organization's data by default"
-
AI data privacy and processing claims
No-training, zero-retention, confidential-data, regional-processing, DPA, and AI privacy wording mapped to buyer questions.
Read guide →Pattern / review use
Data processing · Retention · DPA scope
"Your organization's data always remains confidential, secure, and entirely owned by you"
-
SOC 2 AI claims
SOC 2 Type 2, security controls, audit scope, report period, trust criteria, and AI product-scope claims mapped to buyer questions.
Read guide →Pattern / review use
SOC 2 · Audit scope · AI controls
"SOC 2 Type 2 compliance"
Review and testimonial claims
Authenticity and substantiation standards for AI product reviews and endorsements.
-
Fake AI reviews
AI-generated testimonials, review authenticity signals, and what genuine substantiation looks like.
Read guide →Pattern / review use
Review / Testimonial · Social proof claims
"rate your overall shopping experience so far"
Procurement and vendor evaluation
Buyer checklists for reviewing vendor AI claims before relying on them in a purchase decision.
-
AI vendor questionnaire and DDQ claims
AI vendor questionnaire, DDQ, RFI/RFP, security questionnaire, and AI-CAIQ answers converted into evidence requests and buyer questions.
Read guide →Pattern / review use
Vendor questionnaire · AI-CAIQ · Evidence attachments
"We don't train our models on your organization's data by default"
-
AI vendor due diligence checklist
Pre-purchase checklist for procurement and due-diligence reviewers comparing AI vendor claims, questionnaires, and DDQ answers before contract signature.
Read guide →Pattern / review use
Due diligence · Vendor questionnaire · Before signing
"automatically makes any website fully ADA and WCAG 2.1 compliant"
-
AI claim review boundaries
Route broad AI compliance checker, vendor ranking, procurement approval, and claim verification searches to evidence questions.
Read guide →Pattern / review use
Boundary · Evidence route · Buyer questions
"SOC 2 Type 2 compliance"
Marketing and advertising claims
Marketing-page wording such as superlatives, vague AI labels, compliance references, and outcome claims.
-
High-burden AI marketing wording
Spot high-evidence-burden phrases in public AI marketing copy and map each pattern to evidence buyers should request.
Read guide →Pattern / review use
Vague AI-powered · First / Only / Best · Accuracy / Performance
"the most accurate AI writing detector on the market"
-
ASA/CAP AI advertising claims
When an AI ad uses vague AI-powered wording, no-limits creative claims, or companion positioning, this guide shows what substantiation ASA and CAP expect.
Read guide →Pattern / review use
AI function described · Content restrictions visible · Audience controls · Benefit evidence
"No judgement, no pressure – just a friend who gets you"
Industry-specific claim patterns
SaaS and software claim patterns where AI labels, automation, and compliance wording need closer evidence review.
-
AI washing in SaaS
Claim patterns that indicate AI washing in software products: vague AI features, inflated automation claims, and compliance language with no supporting audit.
Read guide →Pattern / review use
Vague AI-powered · Automation / Replacement · Compliance / Safety
"AI-powered platform — no manual work required"
Have exact vendor wording? Paste the URL or text into the AI product claims checker.
Check the exact AI product claim →