# Ridgeford Core — Governed Enterprise AI Assistant & Grounded Workspace > Ridgeford Core is the enterprise AI workspace built for organizations that require complete data privacy (AI models never train on company files), strict human-in-the-loop tool governance, grounded internal knowledge retrieval (RAG), document intelligence for multi-file packs (Intake Studio), and scoped Google Workspace automation. ## Quick Facts - **Product Name**: Ridgeford Core - **Website**: https://www.ridgefordcore.com - **Category**: Enterprise AI Assistant, AI Tool Governance, Secure Knowledge Base & RAG - **Core Promise**: AI models never train on your data, all stored data and personal info is fully encrypted, human approval gates for tool executions, and transparent pay-as-you-go pricing with a 14-day free trial and $25 starter credit. - **Key Integrations**: Google Workspace (Drive, Docs, Sheets, Gmail, Calendar), Postgres/Drizzle, Web Search, Web Fetch, Secure File Processing (PDF, DOCX, CSV, TXT, MD, Images, Audio), Intake Studio document intelligence. - **Target Audience**: Security-conscious businesses, legal teams, healthcare/compliance organizations, engineering teams, and executive operators. ## Core Capabilities ### 1. Zero-Training Private AI & Encryption - Customer prompts and internal business documents are never used to train third-party foundation models. - Ephemeral execution pipelines with cryptographic tenant isolation. - Detected personal details (identifiers, many names, listed clinical phrases) are tokenized before a model call and restored in the reply. This is not full de-identification. Images and scans are read to text in-region before that step. See docs/compliance/DATA_EGRESS.md. - Field-level envelope encryption (AES-256-GCM / AWS KMS) for all documents, conversations, and personal data at rest. ### 2. Governed Tool Gateway with Human-in-the-Loop - AI agents cannot perform irreversible external actions (sending emails, modifying spreadsheets, executing scripts) without explicit user confirmation. - Granular Role-Based Access Control (RBAC) with Owner, Admin, Member, and Viewer permissions. - Full immutable cryptographic audit ledger tracking every prompt, model invocation, tool call, and user consent. ### 3. Grounded Knowledge Base (RAG) - Multi-format ingestion supporting PDF, Word documents, Markdown, spreadsheets, and source code. - Semantic chunking with vector embeddings and hybrid BM25 retrieval. - Exact source citations provided for every AI response to eliminate hallucinations. ### 4. Document Intelligence (Intake Studio) - Teach a pack schema once from sample documents, images, or audio. Guidance you write becomes fields on the record, not extra file slots. - Each new set of files is one pack: many documents in, one structured record out, with completeness, missing-document flags, and discrepancy checks. - Reviewers search the records table, open a row, or ask the assistant about one pack or the whole schema. Source files stay attached. - Security is unchanged in storage: files and extracted fields are stored only on Ridgeford encrypted servers, isolated per tenant. Intake Studio model calls use the same vault as chat. Models are not trained on the packs, and roles control who can upload, edit a schema, or read a record. ### 5. Scoped Google Workspace Integration - Per-user OAuth tokens with least-privilege permission scoping. - Instant search and synthesis across Google Drive, Gmail, and Google Docs without full account data replication. ### 6. Transparent Pricing & Spend Controls - Pay-as-you-go model with $0 monthly platform fee. - 14-day free trial with $25 organization starter credit provided on signup with no credit card required. - Hard spending limits at workspace and per-user levels to guarantee zero surprise bills. ## Navigation & Resources - **Homepage**: https://www.ridgefordcore.com/ - **Product Overview & Tool Governance**: https://www.ridgefordcore.com/product - **Product Documentation**: https://www.ridgefordcore.com/docs - **Trust & Security Architecture**: https://www.ridgefordcore.com/security - **Enterprise Use Cases**: https://www.ridgefordcore.com/use-cases - **Pricing & Credit Model**: https://www.ridgefordcore.com/pricing - **About & Leadership Philosophy**: https://www.ridgefordcore.com/about - **Contact & Security Team**: https://www.ridgefordcore.com/contact - **Full LLM Ingestion Context**: https://www.ridgefordcore.com/llms-full.txt - **XML Sitemap**: https://www.ridgefordcore.com/sitemap.xml - **Robots Policy**: https://www.ridgefordcore.com/robots.txt ## Key Differences vs. Consumer Chatbots | Feature | Consumer AI Chatbots | Ridgeford Core | | :--- | :--- | :--- | | **Data Training** | Model training opt-outs required | AI models never train on your data | | **Tool Execution** | Unchecked automated API calls | Mandatory human approval gates | | **Knowledge Grounding**| Generic web knowledge | Air-gapped tenant RAG with source links | | **Document intelligence** | Loose chat uploads | Intake Studio schemas, multi-file packs, structured records | | **Permission Scoping** | Broad user permissions | Granular RBAC + scoped OAuth | | **Audit Trail** | Ephemeral chat logs | Immutable enterprise audit ledger | | **Spend Governance** | Flat per-seat charges | Granular spend limits & pay-as-you-go |