About IFS.ai Work Order Generator

The IFS.ai Work Order Generator uses generative AI to automate the creation of work orders by analyzing and interpreting data from an uploaded document, an attached document from the repository, or entered text. It uses an existing model to extract data from the document and a Large Language Model (LLM) to carry out the process. Maintenance engineers often need to reference prior work — documents, spare parts used, and completed tasks — when planning new work, and the previous process lacked a seamless way to link new Work Orders or Work Tasks with that historical work. In practice, you can upload a PDF document, attach an existing document from the Document Management (DocMan) repository, or enter text manually, optionally link the new work to a source Work Order and Work Task, review the AI-generated work order suggestions, and create the work orders directly from the assistant. This improves efficiency and traceability when managing repeat or related work, so important context is not lost. Output quality depends on the clarity, format, and completeness of the uploaded document or entered text.

Who This Feature Is For and Why It Matters

This feature is designed for roles involved in maintenance and service work planning, including:

Maintenance engineers and technicians creating work orders from inspection reports
Service planners and coordinators converting findings into actionable work
Supervisors capturing retrospective work that has already been completed or commenced

Use this feature when you need to:

Create work orders from an inspection report or similar document
Capture retrospective work that is completed or commenced
Reference prior work when planning new or follow-up work
Reduce the manual effort of entering work orders one by one

By using this feature, you can:

Automatically extract suggested work orders from a document or text
Provide information flexibly by uploading a document, attaching an existing repository document, or entering text manually
Instantly reference past work orders and their related documents, resources, and materials for better-informed planning
Link new Work Orders or Work Tasks to a source reference to handle follow-up and recurring work smoothly
Review, edit, and adjust all suggestions before creating them
Generate both new work orders and retrospective work in a single flow

Where to Find It

Access this feature from the IFS.ai Work Order Generator assistant.

Functional Area: Work Order Management

Location: The assistant guides you through three steps — Upload Document or Enter Text, Analyze and Review, and Generate Work. In addition, the Generate Work Order command lets you create a Work Order directly from an existing Work Task, with the source information prefilled.

How the Feature Uses Data

The feature reads the information you provide through an uploaded document, an attached repository document, or manually entered text, and uses it to generate work order suggestions. It can also use a selected source reference to link the generated work to previously completed work.

The feature uses:

An uploaded document in PDF format (only a single file can be uploaded)
An existing document attached from the Document Management (DocMan) repository, specifically a PDF file
Text entered manually into the assistant
An optional source reference from an existing Work Order or Work Task

To use the full functionality:

The uploaded document must be in PDF format and only one file can be uploaded.
The uploaded document or entered text must contain clear and legible information about the work done or to be done.
Poorly formatted documents may be difficult to recognize.
Documents with insufficient data should be avoided.
It is recommended to upload only those pages that contain essential information.
For optimal recognition results, it is recommended to upload documents in English, as the system has been tested primarily with English-language documents. Performance with other languages has not been evaluated.
It is recommended to use ISO date format (YYYY-MM-DD) when providing date-related information, as this format aligns with how the LLM processes and returns date values. This includes a four-digit year, two-digit month (with leading zero), and two-digit day.

Understanding the Output

The feature generates an AI Work Order Report Summary together with suggested work orders extracted from the information provided.

The AI Work Order Report Summary includes the Date of Report, Owned By, Overview, Key Findings, Recommendations, and Prioritization.

The feature also produces two lists extracted from the provided information:

AI Generated New Work Order Suggestions — work to be done.
AI Generated Retrospective Work Suggestions — work that is completed or commenced.

From these lists, you can select the relevant work orders to move on to the next step.

When a source reference is chosen, the Source ID and other details are auto-populated in the Additional Information section of the generated Work Order and Work Task, allowing you to link back to previously completed work.

Enabling and Using the Feature

Prerequisites

AI Services must be enabled.
Sufficient AI tokens must be available.
Every work order must have a document attached. To attach a document at the Work Order level for newly generated work orders, the Object Connection Transformation page must have the following transformations active:
- Target LU: ActiveSeperate
- Service Name: DocReferenceObject
- Source LU: WorkOrder
Similarly, to attach a document at the Work Task level for generated Retrospective Work, the following parameters must be active:
- Target LU: JtTask
- Service Name: DocReferenceObject
- Source LU: WorkOrder

Generate — Step 1: Upload Document or Enter Text

There are three ways of providing information to the assistant: Upload Document, Attach Existing Document from the Document Management (DocMan) repository (PDF files), and Enter Text Manually. You can also select a source reference from an existing Work Order or Work Task so that the generated work is linked back to it. Select Next to proceed to the analysis and review step.

Review and Refine — Step 2: Analyze and Review

An AI Work Order Report Summary is generated, along with the AI Generated New Work Order Suggestions and AI Generated Retrospective Work Suggestions lists. Retrospective Work is the work that is completed or commenced. Review and select the relevant work orders, and confirm that all details are accurate and complete. If the generated work orders are not satisfactory, use Regenerate Work Order to regenerate the work order list with the necessary changes. Select Next to proceed to creating work orders.

Apply — Step 3: Generate Work

The Work Order List and Retrospective Work List contain the work orders that were selected. The extracted list items can be edited in this step to adjust any inaccurate information, and additional work orders or retrospective work can be added. Use the Finish command to generate the work order. The work order is created with the information set.

The document class is determined based on the LU and its class from the Documents Default per Object page. If no value is found, the system checks the Document Default settings, from which it can retrieve the Document Class, Document Revision, and Document Sheet.

You can also use the Generate Work Order command to create a Work Order directly from an existing Work Task. In this case, the source information is prefilled so that the new Work Order remains linked to the originating work.

Example Actions

Based on generated results, you may:

Create new work orders directly from an inspection report
Attach an existing repository document as the basis for new work
Link new work to a source Work Order or Work Task to manage follow-up or recurring work and improve traceability
Record retrospective work that has already been completed or commenced
Edit and add to the suggested work orders before generating them

Improving the clarity, format, and completeness of uploaded documents and text can help improve output relevance over time.

Important to Know

Generated output is a suggestion and must be reviewed before use.

The AI does not guarantee correctness or completeness.

The AI does not replace professional judgment.

Company-specific standards and terminology must be applied manually.

If no results are returned, verify that the input data is clear and complete and that the prerequisites are met.

This model was tested using the following documents and demonstrated high accuracy. You can access the document below to understand the required format when uploading documents or text:

Sample Report

Limitations

Only one source can be processed at a time — one uploaded PDF, one pasted text, or one attached document. Batch or multi-file processing is not supported.
Only PDF files can be uploaded. Word, Excel, and image files are not supported; their content can only be used by pasting it as text.
A run fails if reading, AI analysis, and result creation cannot complete within roughly 60 seconds. There is no fixed maximum file size or page count.
- Feasibility depends on content volume rather than page count: text-dense pages consume more of the window than image-heavy or lightly-filled pages, and a larger number of findings increases processing time.
- Large or finding-dense documents may time out. When the limit is exceeded, the run fails with a timeout or "input limit exceeded" message and must be shortened or split into sections.
Images are not interpreted — only text is read, so a finding shown only in a photo is captured only if the document also includes a caption or description; visual-only defects are missed. In addition, the feature does not create Work Orders automatically; it produces draft suggestions that require manual review and confirmation — especially equipment tags, priorities, and sites — and accuracy is not guaranteed.