Generative AI Build System for Runtime Exception Auto-Healing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing software development systems fail to identify and resolve runtime errors, particularly those caused by version changes in third-party libraries, leading to installation failures that can occur years after product launch, as these errors are not detected during the development process.

Innovation Solution

An intelligent build system using generative AI to predict and auto-heal runtime errors by encapsulating vulnerable code with wrapper code, preserving previous library versions, and redirecting API calls to compatible versions, ensuring seamless execution.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If software applications are built and packaged using conventional build systems, then the build process completes successfully, but runtime errors may occur during installation or execution that were not detected during development

Engineering Contradiction:

Improveruntime error detectionVSAvoidtime to identify runtime errors

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary analysis of build packages using AI models before deployment to predict potential runtime errors. This advance detection allows the system to identify issues during the build stage rather than waiting for runtime, directly addressing the contradiction by improving reliability while reducing the time loss associated with late error detection.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

An AI-based intermediary system is introduced between the conventional build system and the deployment process. This intermediary analyzes build packages, predicts runtime errors, and provides recommendations without replacing the existing build workflow, thereby improving error detection capability while maintaining compatibility with current development timelines.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If AI-based runtime error prediction is implemented, then runtime errors can be identified at build stage, but system complexity increases

Engineering Contradiction:

Improveruntime error predictionVSAvoidbuild system complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The AI-based error prediction system is implemented as a separate, modular component that operates independently from the conventional build system. The build process is segmented into traditional build steps and an optional AI analysis step, allowing organizations to adopt error prediction capabilities without fundamentally complicating their existing build infrastructure.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The AI prediction system serves as an intermediary layer that interfaces with the build system through standardized protocols. This intermediary approach allows runtime error prediction to be integrated without directly modifying the core build system architecture, thereby managing complexity while maintaining prediction capabilities.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Reliability

If comprehensive testing is performed to achieve 100% code coverage, then runtime errors can be detected, but development time and cost increase significantly

Engineering Contradiction:

Improvecode coverageVSAvoiddevelopment speed

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system replaces the mechanical approach of exhaustive manual testing with an AI-based predictive analysis system. Instead of executing comprehensive test suites to achieve 100% code coverage, the AI model analyzes build packages and predicts runtime errors programmatically, maintaining high reliability while preserving development speed.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The AI system performs preliminary error prediction on build packages before deployment, identifying potential runtime issues without requiring exhaustive testing. This preliminary analysis provides a practical alternative to comprehensive testing, maintaining code quality and reliability while avoiding the significant time and resource costs of achieving complete code coverage through traditional testing methods.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12417091B2Auto healing of runtime exceptions using intelligent build system and generative AI

Publication Date: 2025.09.16 BANK OF AMERICA CORP

AI summary

During a software development lifecycle of a software application, application library versions are modified and multiple versions are built and packaged to be installed on different computing systems. An intelligent prediction and capture of runtime errors that may be encountered during a runtime of a software application provides a system-agnostic intelligent software code compiler/build system that intelligently identifies runtime errors at a build stage using artificial intelligence (AI) and automatically encapsulates specific software code to auto-heal runtime exceptions caused by updates to library files. The system may optionally also output a solution to the runtime error to the graphical user interface (GUI) of a software developer capable of performing another software application build.