Home - Main Site (Public User)»Centralpoint AI Governance»AI Interactive Stack
  • Printer Friendly Version
  • Decrease Text Size Increase Text Size

AI Interactive Stack

Centralpoint AI Data Flow Animation

Centralpoint AI Data Flow

Animated Lottie-style plumbing view: data starts from enterprise sources on the left, flows through Centralpoint SQL, Skills, Prompts, Compound Engineering, and GRAIL, then outputs through chatbots to users. Activity flows back into AI 360 to build inertia.

Scheduled daily aggregation

1. Data Sources

365

Microsoft 365

Word, Excel, PowerPoint, Outlook

SP

SharePoint / O365

Lists, pages, libraries, teams content

F

File Paths

Folders, PDFs, documents, media

DB

SQL / Databases

Operational records and rows

{}

APIs / JSON / XML

Structured machine-readable feeds

2. Ingest & Transform

Scheduled Ingestion

Daily pull routines keep the truth fresh.

Normalize + Enrich

De-duplicate, classify, infer, redact, relate.

Save to Index

Each record maps into governed SQL fields.

Centralpoint server

Centralpoint Microsoft SQL Core

System of record where content, metadata, permissions, Skills, Prompts, AI/NLS index, Compound Engineering, and AI 360 activity remain governed inside Centralpoint.

AI / NLS Index

Semantic fields, lexical search, AI vector fields, permissions, and relationships live in SQL.

Skills

Defines what AI is allowed to do, by user, role, channel, and collection.

Prompts

Business logic, instructions, guardrails, and dynamic parameters stay local.

Compound Engineering

Tracks prompt success, user, outcome, and what improves relevance over time.

GRAIL — Governed Retrieval

Security, Skills, Prompts, context filters, and user history narrow the retrieval space before AI is called.

4. AI Model Layer

Embedded / Local LLM

LLAMA, QWEN, ONYX. Runs inside the client environment; data stays local.

OR

External Generative LLM

OpenAI, Claude, Gemini, Copilot. Receives only metered, governed context.

5. Consumption Layer

Centralpoint chatbot

Centralpoint Chatbot

Delivers governed answers through public sites, private portals, apps, Teams, and dashboards.

End User

Receives the right answer, at the right time, in the right channel.

Data ingestion Governed retrieval Chatbot delivery AI 360 feedback loop