Workflow Creation Agent

Select Agent Type

When creating an agent, select "Advanced Agent" (the initial creation steps are the same as for a regular agent).

• Build Type: Two core modes are provided to meet business needs of varying complexity.

  • Advanced Mode: Quickly build complex workflows based on conversations.

  • Application Mode: Create fully orchestrated applications with customized user interfaces and structured input flows, delivering a superior user experience.

Advanced Mode

An enhanced Q&A agent built on top of the basic agent, with core capabilities focused on multi-turn conversations and knowledge-based Q&A.

Features:

• Uses natural language conversation as the main interaction method.

• After creation, it will appear in the Agent List.

• Suitable for business scenarios requiring flexible, open-ended conversational interactions.

Typical Scenarios:

• Internal enterprise knowledge Q&A assistant

• Product FAQ and technical support assistant

• Ticket processing consultation assistant

Application Mode:

Used to build an Agent orchestration application with UI, featuring specific input structures for completing designated tasks. It does not focus on free-form conversation, but rather on executing predefined task flows.

Features:

• After creation, it is displayed on the APP page.

• Has a fixed, customizable input interface (e.g., file upload, form field entry).

• Focuses on executing automated or semi-automated task flows.

• Supports complex input types such as files and structured data, suitable for business scenarios requiring strict input/output formats.

Typical Scenarios:

• Contract review App (upload contract → automatically mark risks)

• Document analysis App (upload document → automatically generate key content)

• Data cleaning App (upload Excel → automatically process data)

Workflow Configuration

Configure the workflow by dragging and connecting the following nodes according to the user's actual business logic:

• Start, End: Built-in input and output modules, customizable input/output parameters and fields;

• Model: Select the model to use in this module, input variables obtained from other modules, and edit prompts and output messages, saving them as variables;

• Knowledge Base Retrieval: Retrieve the most matching information from the selected knowledge base based on input variables and return it;

• Skill: Select one of the skills to perform input/output actions through that skill;

• Code: Customize and create code functions based on output variables from other modules;

• IF: Connect multiple downstream branches, only run the corresponding branch if the set condition is met, otherwise run the "Else" branch;

• Intent Recognition: Used for recognizing user input intent and matching it with preset intent options;

• Text Formatter: Used to process the format of multiple string-type variables;

• Data Source: Select data sources to add referenceable variable content;

• File Retrieval: Search within uploaded files to find relevant answers based on input questions;

• Loop: Repeat a set of tasks for each item in a list, with optional parallel processing.

Node Details

Start

The start node is the entry point of the workflow, used to set the information required to initiate the workflow.

• Function: Preset the basic information (input parameters) required to complete the task. When trigger conditions are met, the system automatically collects and passes these parameters to start the process.

• Processing Logic: By pass; passes input content directly to subsequent nodes.

• Output: All input content.

• Special Note (Application Mode): In this mode, the start node supports defining various complex input types, including file uploads (PDF, Excel, images, etc.) and structured data fields, providing a highly customized input interface for building professional applications.

Model

Invoke large language models to generate responses using variables and prompts.

Input: Select an existing model from the dropdown and choose input variable names. Input Parameters: query (String, from upstream or user input). Configuration Parameters:

• One or more Tools

• Model

  • GPT (GPT or other models)

  • Temperature: controls creativity; the higher the value, the more creative and random the response.

  • Top P: limits the range of selected words by "probability threshold," controlling response diversity.

  • Max Reply Length: limits the maximum number of words the AI can reply at once.

• System Prompt: hidden instructions given to the AI to control overall style

• User Prompt: user input content or questions.

• History: previous conversation rounds to maintain context understanding.

• Processing Logic: Pass the input to the large language model (LLM), which generates responses based on configuration.

• Output : Text content generated by the model.

💡 Tip: You must first connect to a preceding node to select variables from other nodes as input for the current node.

Skill

• Website Reading: Can read static text on web pages (but cannot see dynamically loaded content).

• Text-to-Image : Generates images based on descriptive text and returns the image URL.

• Google Search: Provides real-time, extensive information from the internet.

💡 Tip: In advanced orchestration mode, you can only add built-in skills; custom skills cannot be added.

Code

Write code to process input variables and generate return values.

• Input: Receives externally passed variables, serving as the data entry point required for code execution, providing raw data for subsequent code processing.

• Input Parameters: query (string, code request passed by user or upstream).

• Configuration Parameters: settings related to code execution.

  • Maximum Runtime

  • Code Configuration (Code Input)

• Processing Logic:

  • Runs code in a secure sandbox environment (based on RestrictedPython or designated platform).

  • Limits runtime and access permissions to avoid security risks.

• Output: After code processes the input data, the result is output as a specified variable, serving as the exit for code processing results.

Selector

Acts as a conditional judgment in workflow orchestration. It connects multiple downstream branches and decides the execution path based on set conditions.

• Conditional Branch: Multiple conditions can be set, such as "if - priority 1". By configuring referenced variables, selecting conditions (e.g., equals, greater than, etc.), and comparison values, it judges whether the condition is met. If met, the corresponding branch process runs.

• Processing Logic: Different paths are taken based on different conditions (if none are met, the Else path is taken).

• Output: No direct output, only determines the next node's direction.

Intent Recognition

Intent recognition is a key part of natural language processing. This module analyzes user input content to determine the true intent and match it with preset options.

• Model: Select the model used for intent recognition, which determines the capability and effectiveness of intent recognition.

• Intent Matching: User intent descriptions can be pre-input as matching criteria, and other intents can be added. The system judges which preset intent the user input matches.

• Advanced Settings: Allows setting system prompt content, referencing input variables to optimize prompt effect; also allows setting the number of historical memory entries for the model to reference past conversation information to improve recognition accuracy.

• Processing Logic: Determines the user's true intent and classifies the input into corresponding categories.

Knowledge Retrieval

• Input: Provides retrieval keywords and other raw data for knowledge base search by defining variable names and setting parameter values.

• Processing Logic: Searches the knowledge base based on input and parameters, returning snippets or FAQs.

• Knowledge Retrieval: Selects a specific knowledge base as the search scope; the system searches within this scope for matching information.

• Max Recall Quantity: Set the maximum number of matching results to return from the knowledge base to avoid excessive data.

• Output: Outputs the matched information retrieved from the knowledge base as specified variables for subsequent workflow use.

Text Formatter

Mainly used to process the format of string-type variables.

• Input: Variable names can be defined, and parameter values can be obtained by reference, providing raw string data for subsequent text processing.

• Processing Logic: Simple processing of text.

  • String concatenation

  • String splitting

• String Concatenation: Provides a text editing area where input variables can be referenced by variable names as needed to concatenate multiple strings or perform other formatting.

File Retrieval

File retrieval is a functional module for searching and other operations on file content.

• Input : By defining variable names and referencing parameter values, provides keywords and other input information for file content search.

• File: Files to be processed can be added to this node to define the scope of file retrieval.

Configuration File Retrieval

1. First, you need to check “Support Files” in the “Start” node

2. After checking it, a variable named “USER_FILE” will appear in the Start node, which represents the file provided by the user in the current turn

3. Then, connect the “Start” node to the “File Retrieval” node

4. After the connection, the File Retrieval node can add “USER_FILE” as the file to be retrieved

Data Source

• Data Source: Select the data source to connect.

• Processing Logic: Converts natural language into SQL to query the database and returns results.

• Output: Outputs data from the data source to the next node.

Loop

Used to repeat a set of tasks a specified number of times or over a specified data collection. Different loop modes can be configured to flexibly achieve batch processing or repeated operations.

• Loop Types: Supports two modes

  • Array Loop: Executes tasks sequentially for each element in the input array.

  • Numeric Loop: Executes tasks according to the set number of times.

• Loop Number/Array:

  • When "Numeric Loop" is selected, a specific number must be input, e.g., 2, meaning the task will be executed 2 times.

  • When "Array Loop" is selected, an array variable must be provided; the system will take elements from the array one by one as input to execute tasks.

• Parallel Execution: Optional feature. If enabled, the system processes multiple loop tasks simultaneously to improve efficiency. Users can set the maximum parallel number to control resource usage.

Workflow Example

In this scenario, the workflow function is used to build a complete "Microsoft Support Ticket Issue Analytics" process, with the specific flow as follows:

1. Start Node
   The starting point of the process, included by default in the system.

2. Data Source Node
   Used to connect the raw data needed for ticket analysis.

3. Knowledge Retrieval Node
   Connects knowledge documents containing analysis reference materials as theoretical support for AI analysis.

4. Model Node
   Based on the AI model, combines data source and knowledge base content for comprehensive analysis, generating ticket issue analysis results.

5. End Node
   The endpoint of the process, outputs the analysis results from the model node. This node is included by default in the system.

The data source node and knowledge base node are configured in parallel, while the model node aggregates and processes information from both to ensure the output results have data basis and theoretical support.

Final result:

Note: This example is a simple application of the advanced orchestration feature to demonstrate its basic workflow effect. Advanced orchestration has powerful flexibility and extensibility, supporting complex business logic and intelligent automation workflows through various node types, and can be widely applied in many practical business scenarios.