Ask Industrial AI Assistant questions

Industrial AI Assistant is designed to direct you to specific information in CONNECT visualization in response to the questions you ask. The more relevant a request is, the better the results will be.

Familiarity with the information in CONNECT visualization is therefore helpful when using the assistant.

At the top level, CONNECT visualization presents information across three groups:

• Assets - physical components within an operation. For example, a physical location (a city or site) or a physical object (a pump or a conveyor).

• Streams – the points within an operation that generate data. This can include analog streams, digital streams, time-series streams, alarms and events.

• Content – additional information about an operation or an asset, designed to provide specific visual or statistical analysis. This can include documents, visualizations, charts and graphics.

Operational data is aligned to these groups when you add a namespace from CONNECT data services. Some additional content (such as visualizations) may also be created by users within the visualization services environment.

Industrial AI Assistant can help you access these resources in the following ways:

• Users of AVEVA Connected Manufacture Execution System (MES) can make requests regarding asset efficiency, equipment utilization and other event data.

• Users of AVEVA's document management system Asset Information Management Advanced (AIM-A) can retrieve information from engineering documents that have been marked for inclusion in content searches.

• The assistant can refer to the CONNECT visualization user documentation and provide links to specific topics when required.

You can also ask the assistant to help you create visualizations for the data associated with your assets (see Use Industrial AI Assistant to create visualizations).

When asking the assistant a question, you should consider the following:

• Refer to the names (or aliases) used to define the assets and streams that are available in your operational data.

• Indicate the type of information you require (for example, a list, summarized data, events, documents or a calculation).

• Avoid complicated questions that require the accumulation of data from many assets.

• Use realistic time frames that are defined in relation to the current time (for example, “the last two weeks”).

• If you need information for a specific date (or date range), use the format "15 May 2023". This will avoid any potential ambiguity due to disparate regional settings. Results may be inconsistent if a proper date format is not used (for example, "last Monday" or "14 April this year").

You can use up to 500 characters when asking a question.

Examples

• Can you provide a list of all the wind turbines at Hornsea Windfarm?

• How much power was generated daily by wind turbine GE01 across the last two weeks?

• How much power was generated by wind turbine GE02 from 15 May 2024 to 22 May 2024?

• Can you tell me if any unusual events have occurred with wind turbine GE01 in the last week?

• Can you show me an operational manual for the ongoing maintenance of wind turbine GE01?

• How many hours was the wind turbine GE01 in a fault state last week?

• Show me some KPIs for the Hornsea Windfarm.

• Create a visualization with a column chart that shows the current energy production values for the Hornsea Windfarm.

You can also ask follow-up questions to get additional or related information. If required, the assistant can refer to the conversation history to process any further questions.

• Are there any maintenance manuals for that wind turbine?

• What about for the previous week?

• Is there similar information available for GE02?

• Can you write a summary of the results so far?

There is a limit to the amount of questions Industrial AI Assistant can process within an ongoing conversation, particularly if a complicated response is required. If the assistant reaches this limit, it may ask you to "wait a moment and then make your request again".

Note: Any questions you ask the assistant will be monitored for relevance and responsible usage. If a question is determined to be inappropriate, the assistant will indicate that it cannot provide a response.