This disclaimer explains the nature of EasyPPAP's AI-generated Output and the responsibilities of Customer when using it. By using the Service, Customer agrees to these terms.
1. Why This Disclaimer Exists
EasyPPAP uses generative artificial intelligence to analyze PPAP Material and produce Output. Generative AI is fundamentally probabilistic: it predicts plausible content based on patterns in its training data. It is not a deterministic rule engine and it does not "know" facts in the human sense.
2. What the Output Is
The Output is a decision-support analysis that:
- Identifies likely findings, gaps, or inconsistencies in PPAP Material based on AIAG standards and (when configured) OEM-specific requirements
- Proposes recommended corrections
- Scores submission readiness on a quantitative scale
- Generates an Excel report formatted for SQE workflow
The Output is always probabilistic. The same PPAP Material analyzed twice may produce slightly different findings, especially in borderline cases.
3. What the Output Is NOT
The Output is not:
- A guarantee that the PPAP will be accepted by any OEM, Tier-1, or supplier quality function
- A substitute for the professional judgment of a qualified Supplier Quality Engineer or equivalent reviewer
- A regulatory or legal opinion
- A statement of compliance with any specific industry standard, certification, or regulation
- A replacement for OEM-specific approval workflows
EasyPPAP does not certify PPAPs. EasyPPAP is an analytical aid.
3.1 Scope of the Review
The Service performs a documentary review of the PPAP Material: it verifies presence, formal completeness and cross-document consistency of the elements required for the declared submission level (or, on Level 4, of the documents Customer indicates were requested by its customer), and reads the values stated in the documents against the thresholds defined by the applicable standard or loaded OEM patterns. The Service does not:
- Verify dimensional results measurement by measurement against the drawing, or recalculate statistics (capability indices, gage R&R, etc.) from raw data
- Assess the engineering adequacy of an FMEA, control plan, process flow or any other technical content — only its formal completeness and internal consistency
- Confirm laboratory accreditation, supplier certifications, IMDS status or any other fact against external databases or systems
- Apply customer-specific requirements of OEMs or programs not loaded in the Service, or requirements communicated outside the PPAP Material
- Reliably read handwritten notes, poor-quality scans or content the models cannot extract; such content is reported as unread, never assumed
- Inspect physical samples, master samples or checking aids
Each report includes a Coverage sheet stating, element by element, what was reviewed and what was not. Anything listed as not verified remains Customer's responsibility. The scope described here is also summarized on the Scope of the review section of the website.
4. Mandatory Human Review
Customer commits to:
- Reviewing every Output before relying on it for any decision affecting supplier qualification, production approval, regulatory filing, or contractual obligation
- Verifying findings flagged by the Output against the original PPAP Material
- Treating the Output's confidence scores and recommendations as inputs to the SQE's judgment, not as decisions
Customer's failure to perform mandatory human review does not transfer responsibility to EasyPPAP.
5. Known Limitations
EasyPPAP discloses the following known limitations of the V1 Service:
- Image-only documents: scanned PPAPs without text layers may be partially or fully illegible to the AI. The Output will indicate when image extraction failed.
- Non-standard languages: PPAP content in languages other than English, Spanish, Portuguese, or German is processed less reliably.
- Custom OEM standards: if your OEM uses a proprietary standard not in EasyPPAP's calibration set, findings may be miscalibrated.
- CAD files: EasyPPAP does not analyze geometric tolerances or 3D CAD geometry. CAD files referenced in PPAP Material are not processed beyond bibliographic checks.
- Hallucinations: the AI may occasionally produce findings that reference parts of the PPAP that do not exist or misattribute statements. This is a known property of LLMs and is mitigated but not eliminated by EasyPPAP's retrieval and citation safeguards.
6. Prohibited Uses
Customer may not use the Service to:
- Make or substantiate safety claims about a part or process without independent qualified review
- Generate documentation intended to deceive an OEM, regulator, or supplier
- Substitute for legally mandated human review where law or contract requires a named qualified individual
- Process content unrelated to PPAP analysis (the Service is not a general-purpose AI assistant)
- Process classified, export-controlled, or ITAR/EAR material without Customer's separate confirmation that the relevant sub-processors are authorized recipients
- Use the Output as evidence of compliance in regulatory submissions where the regulator requires direct human authorship
7. EU AI Act Compliance
EasyPPAP monitors developments under Regulation (EU) 2024/1689 (the AI Act). Based on current understanding, the Service is a general-purpose AI tool used for industrial decision support and is not a "high-risk" AI system under Annex III. EasyPPAP commits to maintaining transparency about AI use (this disclaimer), logging Service interactions for incident response, and cooperating with reasonable requests from EU regulators.
If EasyPPAP determines that the Service falls within a high-risk category in the future, EasyPPAP will implement applicable Article 9-15 obligations and notify Customers of material changes.
8. Reporting AI Errors
Customer is encouraged to report errors, hallucinations, or miscalibrations via the in-product feedback mechanism or by emailing admin@easyppap.com. EasyPPAP uses anonymized error reports to improve the Service.
9. Acknowledgement
By using the Service, Customer acknowledges that Customer has read this AI Disclaimer, understands the probabilistic nature of AI Output, and accepts responsibility for human review of Output before reliance.
⚠️ Draft notice: This AI Disclaimer is a draft pending review by qualified US SaaS counsel. EU AI Act enforcement timeline is staggered through 2025-2027; the current classification statement above is provisional. Final language will be confirmed with counsel before public launch.