The description of First Time Fix Prediction is divided into the following sections:
First Time Fix Prediction evaluates whether a Request Task will be resolved successfully without incomplete attempts.
A request task is considered First-Time Fixed when:
Predictions and associated insights can be viewed on:
To activate predictions, train the AI model under Solution Manager / Automation and Optimization / Machine Learning / Trainable Model. After training, set the model to Active to start using predictions. Regular retraining ensures predictions reflect the latest data.
You can schedule prediction jobs in Database Tasks using First Time Fix Batch Prediction with custom intervals to fit your operations. The model supports batch processing and offers flexible configuration to tailor the limit of the allocated start of the Request Tasks considered, in addition to the Company, Service Organization, and Service Delivery Unit applicable.
The model evaluates several key inputs, including:
Training Example: The following example illustrates how various operational, asset-related, scheduling, and capability attributes are captured for a Request Task and used as part of the First Time Fix Prediction model. These attributes help describe the task’s context, highlighting factors such as asset age, historical repeat behavior, material readiness, technician suitability, and scheduling conditions. By combining these elements, the system learns how real-world task characteristics influence the likelihood of achieving a First-Time Fix. The attribute values shown here were recorded for a completed task and contributed to the model’s understanding of task outcome
The following attributes were captured for this task:
Prediction Example: The AI model uses the input attributes to predict the likelihood of resolving the task without incomplete attempts.
For a new incoming Request Task with similar characteristics, the model evaluates several contextual factors such as allocation balance, asset condition, historical behavior, technician capability, and material readiness. In this example, the task shows moderate under-allocation, involves a newer asset, has minor repeat history, benefits from a good skill match, and has materials sufficiently available.
The model may return:
This indicates:
A user-defined FTF Probability Risk Value can be configured in Service Ai Basic Data page. When the First Time Fix Prediction (FTFP) score exceeds this risk value, it is displayed in green, indicating a low-risk and high probability of a successful first-time fix. If the FTFP score falls within or below the defined risk value, it is shown in red and marked as “At Risk”, helping planners quickly identify tasks that may require additional attention. If no risk value is configured by the user, the system applies a default value of 60%.
Tokens are consumed when:
To ensure accurate and functional predictions, the following data and system conditions must be in place:
Prediction requires the First Time Fix Prediction model to be trained with sufficient historical Request Task data ensuring the system can generate reliable probability scores for new tasks.
Top Contributing Factors explain why the model generated a specific prediction. Using the model’s explainability (XAI) layer together with operational context, the system identifies which attributes most strongly affect the likelihood of achieving a First-Time Fix
Typical contributing factors may relate to various aspects of the task, asset, technician capability, materials, location, and scheduling. These attributes influence the likelihood of achieving a First-Time Fix and help explain why the model produces a specific prediction.
These factors appear alongside the prediction score within the On-Page Assistant, enabling planners to interpret the model’s reasoning and improve decision-making.
The recommendation engine provides contextual and actionable suggestions to improve First-time Fix outcomes. Using both the model’s explainability (XAI) and operational context, the system prioritizes recommendations based on impact.
Recommendations may include improvements related to:
Recommendations are accessible through the On-Page Assistant in:
Action Automation enables planners and dispatchers to instantly execute recommended corrective steps, reducing manual effort and improving consistency across teams. Available automated actions include
These automated actions streamline workflows and support standardized best practices.
The Post Completion Validation Window acts as a secondary quality check to ensure that a completed Request Task truly represents a successful fix. It verifies whether any follow-up work is created shortly after completion, helping prevent false First-Time Fix classifications and improving the accuracy of training data used by the prediction model. It works as follows:
Using First Time Fix Prediction, you can: