AI Manage Rules

This module manages structured AI rule records that feed your training pipeline (including vector-style ingestion and embedding workflows).

• Base URL (Admin): https://admin.acusolo.net/

AI Rule list

Under AI Training → Manage Rules.

Columns

Column	Definition
Actions	Edit (update rule) and Delete (remove rule, with confirmation).
Name	Rule name / internal subject of the training item.
Category	Optional grouping label (domain, feature, or topic bucket).
Status	Active state of the rule (for serving/training selection).
Embedding Status	Processing result for embedding pipeline (for example COMPLETED).
Priority	Rule priority level (for ranking/selection behavior).
Files	Number of attached files ingested for this rule (for example PDF count).
Chunks	Number of generated chunks after document split/index stage.
Timestamp	Created / updated dates.

List behavior

• Search by name / category / status.
• Sort and pagination work like other admin tables.
• New AI Rule opens the create modal.

Create AI Rule

Main fields

• Rule Name (required)
• Category (optional)
• Priority (for example Low / High)
• Kind (for example Knowledge)
• Content Method
  • Text Description
  • Upload File
• Description (required content body when using text route)

Scope targeting

• Audience: Global / Male / Female
• Age: Under 18 / 18 to 40 / Over 40
• Language: multiple language chips (fr, en, ar, bn, es, it, de, pt, ja, etc.)

If no scope is selected, UI indicates fallback to GLOBAL.

Advanced fields

• Guardrail Rule (JSON, optional)
• Tags (comma-separated)
• Metadata (JSON-style object)

Actions

• Cancel
• Create AI Rule

Edit AI Rule

Edit opens with existing values and processing summary:

• Status
• Embeddings
• Files
• Chunks

File-based ingestion behavior

When Upload File is used:

• Current files are shown (example: PDF).
• New file upload/dropzone is available.
• Accepted formats and size limit are displayed in the UI (example shows PDF/DOC/DOCX/TXT up to 10 MB).

Update actions

• Cancel
• Update AI Rule

Data and training intent

This module is intended for AI knowledge input where content can come from:

• structured text,
• uploaded documents,
• JSON-style guardrails/metadata.

Configured records are then prepared for embedding/index workflows (shown by Embedding Status, Files, and Chunks) so they can be consumed by your AI training/serving pipeline.