Updated and Effective Date: December 17, 2025
BenchPrep is committed to deploying Artificial Intelligence (AI) in a way that is safe, secure, transparent, and aligned with responsible innovation principles. This document provides answers to frequently asked questions regarding how our AI features (“BenchPrep AI Features”) are designed, governed, and operated within the BenchPrep platform.
All capitalized terms not defined here have the meanings assigned in the BenchPrep Master Subscription Agreement or governing customer contract. BenchPrep uses a combination of Retrieval-Augmented Generation (RAG), vector search, and enterprise-grade large language models (LLMs) to enhance learning experiences. Customer Data remains isolated to each tenant and is never used to train shared models.
What types of AI technologies are used in BenchPrep’s AI Engine?
BenchPrep AI features use:
- Retrieval-Augmented Generation (RAG)
- Natural Language Processing (NLP)
- Supervised learning-based vector similarity search
- Generative AI via third-party LLMs with zero-data-retention and no training on BenchPrep inputs.
Models are sourced from reputable enterprise AI providers and open-source ecosystems, following strict licensing and security evaluation.
Is the AI solution a product, service, or in-house development?
Most of our AI solution is based on supplemental learning with some generative components comprising:
- BenchPrep-developed RAG and vector search architecture BenchPrep UI, workflows, and orchestration layers
- Third-party enterprise LLMs used strictly for inference
- Optional supplemental generative tools integrated into the platform
What are the ethical and legal risks, and how are they mitigated?
BenchPrep actively mitigates ethical and legal risks through:
- Privacy-by-design architecture ensuring tenant isolation and minimizing exposure
- Zero-retention model selection to prevent providers from storing or training on Customer Data
- Bias mitigation procedures applied to prompt engineering and retrieval inputs
- Clear user guidance on appropriate use of AI outputs
- Compliance with IP, copyright, and data protection laws
In using the BenchPrep AI Features, your Usage and Customer Data (including inputs and outputs):
1. are NOT available to other customers.
2. are NOT used to train, re-train, or fine-tune any underlying LLM Models both in-house or Third Party.
What type of data is used by BenchPrep AI?
BenchPrep AI uses only tenant-provided course content and user queries, all of which remain fully isolated within each customer tenant. No data is shared across tenants, and Customer Data is never used to train shared models.
How does the AI system process personal data?
AI features do not use or require personally identifiable information (PII). PII is excluded from vectorization and prompts, and all processing remains covered under existing data protection agreements. Any future PII-related features would require explicit customer opt-in.
How does BenchPrep ensure data security and privacy?
Customer Data stays within BenchPrep’s secure, isolated environment and is encrypted in transit and at rest. Access is controlled through strict role-based permissions, and BenchPrep uses zero-retention enterprise AI providers. Ongoing risk assessments and privacy-by-design practices reinforce data protection throughout the AI lifecycle.
What level of transparency and interpretability is available?
BenchPrep provides transparency into how its RAG components retrieve and use tenant content, while third-party LLMs operate as black boxes. Guardrails and structured prompts help ensure predictable outputs, and additional interpretability features are planned.
What is the liability framework for AI-generated materials?
BenchPrep’s liability commitments are governed by the customer’s Master Subscription Agreement established during procurement.
Who can customers contact for any further questions or concerns about AI trust and safety at BenchPrep?
If you have any further questions or feedback, please do not hesitate to contact your account team, or email us at help@benchprep.com
