AI Agents — Documentation Index

Audience: End users, administrators, developers, and AI agents reading this documentation through the enezadocs MCP system. Use this document when: You need to choose the right AI guide, understand the terminology, or find the canonical document for a topic. Canonical for: Audience routing, terminology, and the documentation map. Related: End User Guide, Administrator Guide, Custom Template Guide, Agent Instructions, Report Design Recommendations, Document Processing.

Enezaframework ships an AI subsystem in the enezaagents app. It provides a contextual assistant on records and lists, a multi-agent AI Workspace, AI audit screens, and document intelligence.

This series explains how to use the subsystem, how to configure it, how to embed it in your own Django templates, and how an AI agent should read the documentation.

Documentation map

Position	Guide	Audience	Use when
0	AI Agents — Documentation Index	All	You need the map, terminology, or reading order.
1	End User Guide	End users	You use the Ask assistant launcher, the AI Workspace, proposals, or documents.
2	Administrator Guide	System administrators, Agent Administrators, AI Auditors	You configure providers, models, tools, agents, versions, permissions, document intelligence, audit, or settings.
3	Custom Template Guide	Developers and administrators	You add Agent Assistance to a custom Django template outside the built-in Enezaframework templates.
4	Agent Instructions	AI agents	You answer questions from this documentation through MCP.
5	Dashboard Builder	Dashboard Creators, end users, administrators, and developers	You want AI to create a dashboard draft, reuse a seeded sample, or explain dashboard-builder behavior.
6	Report Design Recommendations	End users and administrators	You want AI document-type identification and corporate design recommendations for new or edited reports.
7	Document Processing	End users, administrators, developers, and AI agents	You upload documents, configure document schemas or templates, run a document operation, or diagnose a document job.

Who should read which guide

You are	Start here	Then read
An end user	This page	End User Guide
An end user working with documents	This page	Document Processing
An administrator configuring document schemas or templates	This page	Document Processing
A Dashboard Creator	This page	Dashboard Builder
A report or document designer	This page	Report Design Recommendations
A system administrator or Agent Administrator	This page	Administrator Guide
An AI Auditor	This page	Administrator Guide — AI Audit and AI Insights
A developer adding AI to a custom page	This page	Custom Template Guide
An AI agent using MCP	This page	Agent Instructions, then the guide for the current audience

MCP reading order

The enezadocs MCP system can read the whole docs/ tree. Use this AI Agents series as the canonical route for AI topics.

When this documentation is read through MCP, use this order:

Read docs/Documentions/AI Agents/index.md.
Read docs/Documentions/AI Agents/ai-agents-agent-instructions.md.
Read the guide for the current audience: - End-user behavior: ai-agents-end-user-guide.md. - Configuration, permissions, publishing, audit, or settings: ai-agents-admin-guide.md. - Custom template embedding or app registration: ai-agents-custom-template-guide.md. - Dashboard drafts or the Dashboard Builder: ai-agents-dashboard-builder-guide.md. - Report or document design recommendations: ai-agents-report-design-recommendations.md. - Document schemas, templates, extraction, analysis, or jobs: ai-agents-document-processing-guide.md.
Read a related series only when the question crosses that subsystem: - Workflow: docs/Documentions/Workflow/index.md. - Email templates: docs/Documentions/Email Templates/index.md. - Sequence generation: docs/Documentions/Sequence Generation/index.md.

The most specific guide is canonical:

End-user behavior → End User Guide.
Configuration and permissions → Administrator Guide.
Embedding in a custom template → Custom Template Guide.
Dashboard drafts and the Dashboard Builder → Dashboard Builder.
Report or document design recommendations → Report Design Recommendations.
Document schemas, templates, extraction, analysis, or jobs → Document Processing.

If two guides disagree, the more specific guide wins. Do not invent behavior that is absent from the guides.

The four AI surfaces

Surface	Where it appears	What it does
Agent Assistance	Ask assistant on list and detail pages, and on custom pages that embed it	Answers questions about the current view or record, explains related data, and proposes changes.
AI Workspace	Agents › AI Workspace	Runs multi-step chats and tasks with one or more agents, supports documents, and reviews proposed changes.
AI Audit	Agents › AI Audit	Read-only audit of conversations, messages, runs, document jobs, and audit events.
AI Insights	Agents › AI Insights	Dashboards for personal, auditor, and administrator AI usage.

Terminology

Term	Meaning
Agent	A configured AI worker with a name, code, instructions, model, tools, and optional sub-agents.
Agent Version	A snapshot of an agent's instructions, model configuration, tools, and sub-agents.
Published version	The version the runtime is allowed to execute. Draft versions do not run.
Tool	A capability the agent may call, such as metadata lookup, record query, document analysis, workflow inspection, or a developer handler.
LLM Provider	The external model service and its connection settings.
Model Configuration	The provider plus model name, generation settings, capabilities, and cost rates used by an agent version.
Contextual Assistant	The built-in agent used by Ask assistant on record and list pages.
Dashboard Builder	The built-in agent that creates dashboard drafts from a description on dashboard pages.
Report Builder	The built-in agent that creates report drafts and applies design recommendations on report pages.
Report Designer Assistant	The page-specific mode that proposes design changes to the saved report open in the report designer.
Workspace Orchestrator	The permanent agent in the AI Workspace that can coordinate specialists and borrow non-selected specialist tools.
Proposed action	A change the AI prepares but does not apply. A user must accept or reject it, and apply it, before data changes.
AI Workspace	The chat-centric multi-agent workspace under the Agents menu.
Document Analysis Template	Reusable instructions and an output schema for analysing documents.
Document Extraction Schema	A definition of fields, rules, and evidence for extracting data from documents.
Document Processing Job	A recorded document operation with status, results, warnings, errors, and evidence.
Document Category	A classification label stored on a document and optionally used by a schema for lookup.
Evidence	The page, sheet, excerpt, offsets, confidence, and method that support an extracted value or finding.
Application Registration	A code-level registration that tells the AI subsystem which application and models it may describe and query.
Model Registration	A code-level registration for one model, including its business purpose and field policies.

What the documentation covers

How end users ask questions, read answers, review proposals, and work with documents.
How administrators configure providers, models, tools, agents, versions, publishing, permissions, audit, and settings.
How developers add Agent Assistance to a custom Django template.
How the Dashboard Builder creates dashboard drafts from descriptions and seeded samples.
How the report builder and report designer identify document types and recommend corporate designs.
How documents are parsed, classified, extracted, analysed, compared, and recorded as processing jobs with evidence.
How developers register an application, model, field policy, relationship, business action, context provider, or agent from code.
How AI agents should read and cite this documentation through MCP.

A note on testing

Per project rules, the developer or agent does not run or claim test results. The user is solely responsible for all testing and validation of the behavior described in these guides.