Natural Language to Expression Code

The Natural Language to Expression Code feature is an intelligent AI-powered assistant integrated directly within the akaBot Studio Expression Editor. It allows developers to input logic descriptions in plain, natural language to automatically compile and generate precise expression code in VB.Net or C#.

Rather than memorizing complex syntax or spent time searching through programming libraries, developers can describe their intent. The AI assistant acts as an autopilot, accelerating workflow design, enabling low-code/no-code accessibility for beginners, and saving time for experienced developers.

Key Capabilities

The Natural Language to Expression Generator in akaBot Studio offers the following capabilities:

• Converts natural language into VB.Net or C# expressions.
• Helps developers write expressions faster with less manual coding.
• Reduces syntax errors and improves productivity.
• Simplifies creating workflow logic, validations, and calculations.
• Makes expression development easier, even for users with limited programming knowledge.

Configuration

Before using Natural Language to Expression Code, configure the AI provider in Studio.

Open the AI Integration tab

1. Open akaBot Studio.
2. Go to Options (Backstage / Home → Options).
3. Select the AI Integration tab.
4. In Provider, choose one of the following:
   • AI Hub
   • Akabot Cloud AI Server (Cloudflare Worker)

Provider: AI Hub

Use AI Hub when Studio talks to Center through a connected Agent.

1. In AI Integration, set Provider to AI Hub.
2. Open akaBot Agent.
3. Go to Connect and connect the Agent to your Center (same organization / environment used by Studio).
4. Confirm the Agent stays connected while you use expression generation.

If the Agent is not connected or not assigned correctly, Studio cannot authenticate to AI Hub and expression generation will fail.

Provider: Akabot Cloud AI Server (Cloudflare)

Use Akabot Cloud AI Server when Studio calls a Cloudflare Worker endpoint over the internet.

1. In AI Integration, set Provider to Akabot Cloud AI Server.
2. Enter the API Endpoint (Worker URL). This field appears only when this provider is selected.
3. Ensure the PC has internet access to that endpoint.

The API endpoint powers AI features, including expression generation and AkaNinja Chatbot assistance.

NOTE

• AI Hub: Requires a connected Agent to Center. Internet-only access is not enough if Agent is disconnected.
• Akabot Cloud AI Server: Requires a valid Worker URL and outbound internet access.

User Interface (UI) Overview

The feature is embedded directly at the bottom of the standard Expression Editor dialog in akaBot Studio.

1. Expression Area (Editor Pane): The main text field where the generated expression is inserted. You can also manually edit or refine the code here.

2. Natural Language Input Field: The text box at the bottom of the dialog with the placeholder: Describe your expression in natural language.... This is where you enter your prompt.

3. Control Buttons:

• OK: Saves the current expression and applies it to the activity's property.
• Cancel: Closes the dialog without saving.

Step-by-Step Guide

To generate expression code using text prompts, ensure your akaBot Studio AI Integration provider is configured (see Configuration).

To generate expression code using a text prompt, follow these steps:

Step 1: Navigate to the Workflow

Go to the workflow file where you want to add the expression.

NOTE

• Exact Variable Matching: If your expression depends on multiple variables or arguments, your prompt must include each variable or argument name exactly as defined in your workflow.
• Pre-declared Variables: Ensure that all referenced variables and arguments are already created in the Variables or Arguments panels of akaBot Studio before attempting to generate the expression.

Step 2: Open the Expression Editor

Select the activity in your workflow. In the Properties panel on the right, locate the property where you wish to add the expression (e.g., a condition, path, or value). Click the configuration or ellipsis button next to the property field to open the Expression Editor window.

Step 3: Input the Prompt

Locate the input field at the bottom of the Expression Editor dialog with the placeholder: Describe your expression in natural language.... Write a text prompt describing the logic or condition you wish to add.

Step 4: Generate the Expression

Press Enter on your keyboard or click the generate icon. The AI assistant processes your request and populates the generated VB.Net or C# expression code directly in the main editor pane.

Step 5: Save and Apply

Review the generated expression in the editor pane. Make any necessary manual refinements, then click OK to save and apply the expression to the activity.

Best Practices for Writing Prompts

To get the most accurate and reliable expressions from the AI assistant, follow these prompt guidelines:

• Reference Variables Exactly: Always use the exact names of the variables, arguments, or assets defined in your project (e.g., customerEmail, invoiceDateText).
• Specify Action Verbs: Use explicit programming verbs to guide the logic (e.g., "Convert", "Filter", "Extract", "Combine", "Check whether").
• Define Default Values: If handling potentially empty values, specify the fallback in the prompt (e.g., "Convert amountText to Decimal, using 0 if it is empty").
• Break Down Complex Logic: If your logic contains multiple nested conditions, describe them sequentially (e.g., "Check if dt_EmployeeData is not null and has at least one row").

Requirements and Limitations

Keep the following rules in mind to avoid compilation errors:

1. Pre-declare Variables: The AI assistant maps names to variables in your project. Ensure that all referenced variables and arguments are already created in the Variables or Arguments panels of akaBot Studio before generating the expression.
2. Project Language Context: The expression format (VB.Net vs. C#) is determined by your project settings. If your project is a VB.Net project, the editor generates VB expressions; if it is C#, it generates C# expressions.
3. AI provider configured: Configure Options → AI Integration first. For AI Hub, connect Agent to Center. For Akabot Cloud AI Server, set the Worker URL and ensure internet access.

Examples Library

Below is a library of common prompts organized by the editor context in which they are used.

Expression Editor Examples

These prompts are designed for the Expression Editor to generate single-line expressions (e.g., calculations, condition checks, simple string/path manipulations, datatable filtering).

Expression Target	Function	Example Prompt
App variables	Checks if a workflow variable str_UserToken is empty and assigns a default token value.	"If str_UserToken is null or empty, set it to 'DEFAULT_TOKEN'"
App arguments	Calculates the transaction fee and assigns it to the output argument out_TransactionFee.	"Calculate out_TransactionFee as 5 percent of in_Amount"
Filter	Filters a DataTable variable to exclude rows containing invalid or blank email addresses.	"Filter dt_CustomerData where Column 'Email' is not empty and contains '@'"
Custom Model	Retrieves the current URL from the browser variable.	"Get the current URL from edgeBrowser"
Filter	Checks if the DataTable dt_EmployeeData is not null and has at least one row.	"Check if the datatable dt_EmployeeData is not null and has at least one row"
Files	Extracts the file name without extension from a full file path.	"Extract the file name without extension from fullFilePath"
Datetime	Creates a timestamp string in yyyyMMdd_HHmmss format.	"Create a timestamp string in yyyyMMdd_HHmmss format"
Convert	Converts a text representation of an invoice date to a DateTime object.	"Convert invoiceDateText to a DateTime"
Check	Converts a text amount to Decimal, defaulting to 0 if it is empty.	"Convert amountText to Decimal, using 0 if it is empty"

AI Code Editor Examples

These prompts are designed for the Code Editor (used in Invoke Code or Script activities) to generate multi-line logic, complex loops, error handling blocks, database queries, and web service calls.

Expression Target	Function	Example Prompt
App variables	Iterates over input items, cleans the values, and populates an output list.	"Loop through in_RawData list of strings, trim whitespace, convert to uppercase, and add to list_ProcessedData"
App arguments	Maps database query parameters from incoming workflow arguments to a command execution.	"Create SQL command using connectionString and in_Query, bind argument in_UserId, execute query, and assign output to out_ResultDataTable"
Custom model	Downloads web content or communicates with API endpoints within a custom C# or VB.NET script block.	"Create an HttpClient to make a GET request to in_Url, parse the response status code, and output JSON string to out_ResponseJson"
Filter & Aggregation	Grouping, aggregation, and complex joins on data tables that are too verbose for a single-line expression.	"Group dt_Sales by Column 'Region' and compute the sum of 'Revenue' for each region into a new DataTable out_Summary"
Error Handling & Logging	Implements standard try-catch blocks around sensitive file operations to log errors without breaking the main bot thread.	"Try to read file at in_FilePath, catch all exceptions, write details to Console.Error, and set out_IsSuccess to false"
File System Manipulation	Performs deep recursive file scans, filtering, and copying operations.	"Recursively search directory in_SourcePath for all files modified in the last 24 hours, copy them to in_DestPath, and count total files copied in out_FileCount"

Troubleshooting

If you encounter network connection issues or errors when using the Ask AI service, refer to the Troubleshoot Ask AI guide.