AI Studio

Overview

• AI Studio is the AI capability hub of the Bonree ONE platform. It provides unified management of foundational capabilities — Models, Tools, Skills, and Knowledge Base — and supports the full lifecycle of agent development and operations, with Rui AI serving as the unified interaction entry point for all agents.
• The AI Studio capability framework is built on three layers:

Foundation Layer: Models, Tools, Skills, and Knowledge Base — providing the core reasoning, action, and retrieval capabilities that power agents.

Agent Layer: Supports three build modes — Workflow, Autonomous Decision, and Knowledge Planning — covering a wide range of scenarios from fixed-path processes to autonomous multi-step reasoning.

Interaction & Operations Layer: Rui AI serves as the unified user-facing interface; Diagnosis History and Diagnosis Statistics close the operations loop by continuously tracking and improving diagnostic performance.

Value

1. Unified AI foundation — lower the barrier to entry

Models, Tools, Skills, and Knowledge Base are all managed centrally in AI Studio. All feature modules share the same resource pool, eliminating duplicate configuration and resource fragmentation.

The platform ships with built-in models, MCP tools, Skills, and knowledge documents that are ready to use out of the box. Even less experienced users can quickly enable AI-powered fault diagnosis without starting from scratch.

2. Flexible agent building — cover diverse operations scenarios

Three build modes — Workflow, Autonomous Decision, and Knowledge Planning — allow operations teams to choose the most suitable agent type based on the scenario and their team's capabilities.

Agents can simultaneously reference Tools, Skills, and Knowledge Base, enabling end-to-end automation of the "query → reason → execute → notify" pipeline and significantly reducing manual intervention.

3. Knowledge-driven — give AI an understanding of business context

Upload troubleshooting experience, remediation SOPs, and system architecture documentation to the Knowledge Base, enabling AI to understand business context during diagnosis and significantly improving diagnostic accuracy and relevance.

A single knowledge base can be referenced by multiple agents simultaneously — maintain once, apply everywhere, avoiding duplicate effort.

4. Open tool ecosystem — continuously expand diagnostic capability

Integrate custom tools (log queries, metrics analysis, change retrieval, etc.) via the MCP protocol. Combined with built-in tools, this forms a comprehensive diagnostic tool chain.

The Skill system provides action-oriented capabilities such as sending notifications, creating tickets, and pushing logs — enabling agents to not only diagnose but also remediate.

5. Closed-loop operations — data-driven continuous improvement

Diagnosis History captures every execution record, user rating, and negative feedback reason, providing precise problem localization for strategy optimization and tool enhancement.

Diagnosis Statistics presents key operational metrics — trigger volume, success rate, and mode distribution — in visual charts, giving management quantitative visibility into AI diagnostic performance.

Feature Modules

Rui AI

Rui AI is the unified interaction entry point for AI Studio. Users can directly invoke published agents through a conversational interface to complete tasks such as fault diagnosis, intelligent Q&A, and automated reporting. Rui AI supports conversation history, allowing users to switch between sessions and restore context.

Key capabilities:

Conversational invocation: Describe problems in natural language. Agents automatically understand intent and invoke the appropriate tools, knowledge, and models to complete the task.

Multi-agent selection: Select the desired agent or Skill (fault diagnosis, intelligent Q&A, etc.) from the input area and switch flexibly based on the scenario.

Session history: Conversation history is retained for one week and can be resumed to continue in context.

Result presentation: Diagnostic conclusions, charts, and tool invocation traces are clearly displayed in the conversation stream with Markdown rendering support.

Models

The Models module supports integrating large language models that are compatible with the OpenAI API protocol. It is the foundation of all intelligent reasoning capabilities in AI Studio. Integrated models can be selected and used by workflow nodes, Autonomous Decision Agents, knowledge retrieval, and other scenarios.

Supported model types:

LLM (Large Language Model): Used for reasoning, summarization, Q&A, and generating diagnostic conclusions.

Embedding Model: Used for vectorizing and retrieving knowledge base documents. Required when creating a knowledge base.

Reranking Model: Used to improve the precision of knowledge retrieval results. Applied in Knowledge Planning agents.

Key operations: Add / edit / delete models; filter by model type, name, or ID. The system automatically validates connectivity when saving, surfacing integration issues early.

Tools

The Tools module integrates external service tools into the platform via the Model Context Protocol (MCP), making them available for agents to invoke during reasoning. It also provides a built-in set of ready-to-use tools. Tools are the core capability carrier for agents performing "query and analysis" operations.

The Tools module provides two management views:

MCP Services View (default): Manages tools at the MCP Server level, displaying each server's basic information and tool count. This is the primary entry point for adding, editing, and deleting MCP Servers. The built-in MCP Server is always pinned to the top and cannot be deleted.

Tools View: Displays all tools from all MCP Servers in a flat list. Supports filtering by name, description, label, or parent MCP Server, and allows adding labels to tools for categorization and quick lookup.

Key operations: Add custom MCP Servers (with request header configuration and timeout settings); one-click refresh of all server tool lists; click a service card to view complete tool details including parameter descriptions.

Skill

A Skill is a reusable unit encapsulating a specific action capability, enabling agents to not only query and diagnose but also take remediation actions. Skills are either built in by the platform or imported by users from external sources.

Examples of built-in Skills:

Send notifications via DingTalk / WeCom / Email

Automatically create fault tickets in an ITSM system

Push operations logs to a log platform

Skill management features:

Built-in Skills only require parameter configuration (e.g., Webhook URL, SMTP settings) to be invoked in agents — no development needed.

Supports importing external Skill packages (.zip / .skill format) to expand the platform's action capability beyond built-ins.

When configuring an agent, dedicated parameters can be set for each Skill to override global configuration, enabling differentiated invocation.

Knowledge Base

The Knowledge Base is the digital knowledge repository for operations teams. It supports centralizing technical expertise, troubleshooting SOPs, system architecture documentation, and more. After vectorization processing, documents become available for agents to retrieve and reference, giving AI the ability to understand business context.

Document processing pipeline: Upload → Queued → Indexing → Available. Only documents in "Available" status can be referenced by agents.

Key operations:

Select an Embedding Model when creating a knowledge base; multiple document files can be uploaded.

Documents support rename, export (original file), and delete. Batch deletion is also supported.

The Embedding Model cannot be changed once documents have been uploaded to the knowledge base. To use a different model, create a new knowledge base.

The same knowledge base can be referenced by multiple agents (workflow nodes, Autonomous Decision Agents) simultaneously — maintain once, apply to all scenarios.

Agents

Agents are the core capability carriers of AI Studio, encapsulating the reasoning logic, tool invocation chain, and knowledge sources for a specific scenario. The platform provides built-in agents ready to use out of the box, and also supports user-created custom agents.

Build Modes

Build Mode	Core Mechanism	Best For	Typical Configuration
Workflow	Visual node-based orchestration; fixed execution path; LLM can be used as a node	Diagnosis tasks with clearly defined steps and stable processes	Model node + Tool node + Knowledge node + Skill node
Autonomous Decision	ReAct framework; LLM independently decides tool invocations and reasoning rounds	Complex scenarios with unknown root causes requiring multi-round dynamic reasoning	Reasoning model + MCP tools + Skills + associated knowledge
Knowledge Planning	Transforms operations knowledge documents into structured diagnostic workflows; LLM follows the steps	Scenarios that rely on an existing knowledge base and follow a standard troubleshooting path	Reasoning model + reference knowledge document

Agent Status

Draft: Not yet published. Editable, but cannot be invoked in Rui AI or fault diagnosis.

Published: Can be directly invoked in Rui AI and alert-triggered fault diagnosis. Further edits require re-publishing to take effect.

Key Operations

Create / Clone: Completed in a two-step flow (Step 1: Basic setup → Step 2: Build configuration). When cloning, the source agent's configuration is pre-filled automatically.

Edit: Custom agents support full editing. Built-in agents only allow editing a subset of fields, such as the reasoning model and associated knowledge.

Export (Workflow mode only): Export the workflow canvas as a YAML file for backup or migration.

Module Collaboration

All modules in AI Studio work closely together to form a complete AI capability chain:

Relationship	Description
Models → Agents	The reasoning capabilities of agents (workflow nodes, Autonomous Decision) depend on LLMs integrated in the Models module. Knowledge base indexing depends on the Embedding Model.
Tools → Agents	Autonomous Decision Agents invoke MCP tools configured in the Tools module during reasoning to perform log queries, metrics retrieval, change lookups, and other operations.
Skills → Agents	Agents can invoke Skills to perform direct queries or execute remediation actions — such as sending alert notifications or creating tickets — enabling end-to-end automation from detection to resolution.
Knowledge Base → Agents	Workflow knowledge nodes and the "Associated Knowledge" field in Autonomous Decision Agents can reference Knowledge Base documents, providing business context for AI reasoning.
Agents & Skills → Rui AI	Published agents and Skills can be directly invoked by users in Rui AI for interactive fault diagnosis and Q&A.
Agents → Alert Diagnosis	Published agents can be selected in Fault Diagnosis Strategies. When an alert is triggered, the corresponding agent executes automatically to complete the diagnosis.
Diagnosis Execution → Analytics	All alert-triggered diagnosis execution records flow into Diagnosis History and Diagnosis Statistics, forming an operational data loop that drives continuous optimization of strategies and knowledge.

Permissions

The AI Studio permission system covers two dimensions: feature permissions and data permissions.

Feature Permissions

AI Read-Write Access: Full access to all AI Studio pages, including create, delete, edit, and import operations.

AI Read-Only Access: View-only access to all AI Studio pages; supports search and export, but no write operations.

Data Permissions

Models: Environment-level isolation. All resource domains within the same environment share the same model list. The model list changes when switching environments.

Tools, Skills, Knowledge Base, Agents: Resource domain isolation. Content created in one resource domain is not visible in others by default. Content created in the "All" resource domain is visible and usable across all resource domains the account has access to.

Rui AI, Diagnosis History, Diagnosis Statistics: Resource domain isolation. Data is isolated between primary accounts, but shared between a primary account and its sub-accounts, and between sub-accounts. Switching to the "All" resource domain shows data from all resource domains accessible to the current account.

Quick Start

The following is the recommended configuration path for enabling AI Studio fault diagnosis from scratch:

Step	Action	Description
1	Add a model	On the Models page, add at least one LLM. If you plan to use a Knowledge Base, also add an Embedding Model.
2	Configure tools (optional)	On the Tools page, connect the MCP Servers you need. Built-in platform tools require no configuration and are ready to use immediately.
3	Configure Skills (optional)	On the Skill page, fill in the configuration parameters for the built-in Skills you plan to use (e.g., notification channel Webhook URL).
4	Build a knowledge base (optional)	On the Knowledge Base page, create a knowledge base and upload troubleshooting documents. Wait for the document status to reach "Available" before referencing.
5	Create an agent	On the Agents page, create a new agent, select a build mode, configure the model, tools, Skills, and knowledge base, then publish.
6	Configure a diagnosis strategy	In Fault Diagnosis Strategy, configure alert trigger rules to associate published agents with alert filter conditions, enabling automatic diagnosis.
7	Review diagnosis results	View diagnostic conclusions in Rui AI or on the alert detail page. Use Diagnosis History to review past execution records and submit ratings.