AI Assistant

AI Assistant
 provides answers, guidance, and troubleshooting for
FactoryTalk Optix
 projects.

Overview

Interact with
AI Assistant
 in the
AI Assistant
 pane.
AI Assistant pane
AI Assistant pane with a question about events and alarms, a response area, numbered callouts for window controls, the main content area, and the message entry field. The screenshot includes the minimize and close controls, a scroll bar, and the input box at the bottom.
Interface items
Item
Name
Description
1
Clear chat history icon
Clear chat history
 icon
Clear current chat. Useful if you want to start a new chat.
2
Close icon
Close
 icon
Close
AI Assistant
 pane.
To toggle the
AI Assistant
 pane, select
AI Assistant icon
AI Assistant
 from the top bar.
3
Chat
 messages
See your messages and
AI Assistant
 responses.
AI Assistant
 responses may include these sections for complex answers:
  • In Short
    . Summary of the response. Useful if you need a quick answer to your question.
  • Details
    . More information. Useful if you want to deepen your knowledge about a specific subject. May contain links to related documentation and other materials.
4
Enter message
 field
Enter a message to begin or continue a chat with
AI Assistant
.
Send your message either by pressing the
keyboard key or selecting
Send icon
Send message
.
For more information on how to interact with
AI Assistant
, see Use AI Assistant.

Operation

AI Assistant process
AI Assistant design-time assistant receives a user prompt, communicates through the NATS protocol for information search, uses retrieval-augmented generation (RAG) to process the request, and sends a prompt to Rockwell Automation AI services. The response from the LLM returns through NATS, providing the user with an answer.
AI Assistant
 integrates with
FactoryTalk Optix Studio
 as a design-time assistant. The AI is unable to write directly to controllers, change runtime systems, or perform safety-critical actions. The AI does not make changes to your project or system automatically.
The AI communicates with
FactoryTalk Optix Studio
 by using the
NATS
 protocol. The AI uses a large language model (LLM) that
Rockwell Automation
 hosts in the cloud and accesses through secure APIs. The APIs send user messages for processing and return responses.
The AI uses retrieval-augmented generation (RAG) to provide answers based on documentation.
IMPORTANT:
AI Assistant
 may display inaccurate information. Please refer to
FactoryTalk Optix
 documentation to verify its responses.

Use cases

AI Assistant
 helps you find answers about
FactoryTalk Optix
 without searching documentation. Use
AI Assistant
 to access information, learn best practices, and troubleshoot issues in your projects.
Get guidance for common tasks involving
FactoryTalk Optix
:
  • Get step-by-step guidance for tasks such as installing and activating
    FactoryTalk Optix Studio
    , managing projects, processing data from field devices, designing business logic, visualizing data, securing
    FactoryTalk Optix Applications
    , and extending projects with custom C# NetLogics.
  • Ask about the current
    FactoryTalk Optix
     release details.
Expand your knowledge about
FactoryTalk Optix
:
  • Learn best practices for
    FactoryTalk Optix
     projects.
  • Access tutorials and hands-on examples.
Resolve issues with
FactoryTalk Optix
:
  • Find answers to frequently asked questions.
  • Check system requirements.
  • Review known anomalies and find solutions.

Usage limits

AI Assistant
 enforces message, session history, and response size limits. These limits help control request handling and response delivery.
Usage limits
Limit
Value
Maximum user message size
5 KB
Maximum user request history per chat or session
100 requests
Maximum size of all messages per chat
1 MB
Maximum LLM response size
50 KB
If a request exceeds a limit,
AI Assistant
 does not process the request and displays an error message.
AI Assistant
 informs you when LLM responses exceed 50 KB and the system blocks or truncates the responses.
Provide Feedback
Have questions or feedback about this documentation? Please submit your feedback here.
Normal