Challenge by Sidharth Shukla ūüöÄ ERP Testing with AI-Driven Solutions


Userlevel 7
Badge +2
  • Community Manager
  • 105 replies

We are excited to present this month challenge created by @sidharth shukla on how AI can help with challenges in the ERP testing. 

Enterprise Resource Planning (ERP) systems are complex and extensive, making testing a formidable task. The intricacy of these systems presents unique challenges such as test data management, test case prioritisation, and test automation. Your challenge is to devise an Artificial Intelligence (AI)-based solution to address these complexities, thereby enhancing the efficiency and effectiveness of ERP testing.

Background Information:

  1. ERP systems: ERP systems are comprehensive software solutions that integrate and automate diverse business operations, including finance, human resources, supply chain management, and customer relationship management.
  2. ERP testing challenges: The complexity of ERP systems leads to issues such as managing voluminous and varied test data, prioritising test cases for maximum coverage, and automating tests in a resource-efficient manner.
  3. AI in ERP testing: AI technologies like machine learning, natural language processing, and computer vision can revolutionise ERP testing by automating test data management, predicting defects, optimising test coverage, and streamlining test automation.

Challenge Objectives:

  1. Suggest an AI-based solution that addresses the key challenges in ERP testing, including test data management, test case prioritisation, and test automation.
  2. Define an approach that leverages AI to generate relevant test data, predict defects, and prioritise test cases based on their impact on business processes and risk of failure.
  3. Demonstrate how your AI solution improves collaboration between various stakeholders involved in the ERP testing process, thereby enhancing the overall quality and agility of ERP implementation.

Prizes and Points:

ūüŹÜ 2 Winners: Personalized Certificate of Achievement signed by Sidharth, +300 points, a badge, and a gift box from us.

ūüĆü All Participants: +150 points for your valuable contribution to the challenge.

Submission Requirements:

  1. No need to design and develop any solution using AI, rather pick the problem statement and try to come up with a write up on How AI can be used to solve or minimise the efforts. We are specially looking for five  important points: test data management, test case prioritisation, and test automation, predict defects, collaboration between various stakeholders
  2. A proof-of-concept implementation showcasing your solution's key functionalities, such as test data generation, defect prediction, or test case prioritisation.

Evaluation Criteria:

  1. Innovation: The originality and creativity of your approach in leveraging AI to address ERP testing challenges.
  2. Practicality: The feasibility of implementing your solution in real-world ERP testing scenarios, considering constraints like cost, time, and resources.
  3. Impact: The potential of your solution to significantly improve the efficiency and effectiveness of ERP testing and its applicability across various industries and ERP systems.

Key Dates:

  • Challenge release: May 8
  • Lasts: 3¬†weeks
  • Judging (for ShiftSync members): 3 days
  • Winners announced: June 4

Example

AI can analyse historical data and generate synthetic data that accurately represents real-life scenarios, reducing the time and effort required for test data management in ERP systems.
Imagine a company that uses an Enterprise Resource Planning (ERP) system to manage its business operations. As part of their testing process, they need to generate large amounts of test data to simulate various scenarios, such as inventory management, sales transactions, and financial reporting.

Traditionally, creating this test data manually can be time-consuming and labor-intensive. However, by leveraging AI technology, the company can streamline this process significantly.

Using AI algorithms, the system can analyze historical data from the ERP system and identify patterns, trends, and correlations. It can then use this information to generate synthetic data that accurately represents real-life scenarios.

For example, if the company wants to simulate a busy sales day with high transaction volumes, the AI can generate synthetic sales orders, customer data, and inventory records based on past sales data. This synthetic data closely mimics the characteristics of real data, allowing the company to perform thorough testing without the need to manually create or manipulate test data.

By automating the generation of test data using AI, the company can save time and effort, improve the efficiency of their testing process, and ensure more comprehensive test coverage in their ERP system.

Disclaimer: The use of generative AI to solve this challenge, is allowed in the context of formulation but NOT ideation. Every answer should have a clear "human element" otherwise said answer will be disqualified. 

Let the Challenge Begin!

 


6 replies

Userlevel 1

Ace ERP Testing with "EvoTest": Your AI-Powered Evolutionary Solution

 

ERP testing can be a daunting task. But what if you had an AI assistant that constantly learns and adapts, evolving alongside your testing needs? EvoTest is that solution.

Here's how it tackles the five key challenges:

 

1. Self-Evolving Test Data Management:-

 

Adaptive Data Synthesis with Reinforcement Learning: EvoTest utilizes cutting-edge reinforcement learning to analyze test results and user feedback. This allows it to continuously learn and improve the data generation process. Imagine EvoTest creating increasingly realistic test scenarios based on past test failures, ensuring edge cases are constantly covered.

Dynamic Data Augmentation: EvoTest doesn't just generate data. It utilizes AI to dynamically augment existing data with realistic variations and anomalies. This includes simulating unexpected user behavior, network disruptions, and other potential challenges, ensuring comprehensive test coverage.

 

2. Predictive Defect Identification with Explainability:-

 

Multi-source Defect Prediction: EvoTest goes beyond historical data. It incorporates diverse sources like code analysis, user feedback, and real-time system logs for a comprehensive view. This multi-source approach improves the accuracy of defect prediction.

Explainable AI for Prioritization: EvoTest doesn't just predict defects; it explains them. The "Why" behind the prediction empowers testers to prioritize effectively. Imagine EvoTest highlighting a specific line of code as a potential cause of a predicted defect, allowing developers to address it directly.

 

3. AI-powered Test Automation Suite:-

 

NLP-driven Smart Test Case Generation: EvoTest utilizes advanced NLP to analyze not just user stories and requirements but also past test case successes and failures. This allows it to generate highly efficient test scripts that target potential weaknesses, maximizing test coverage.

Cognitive Test Automation Framework: Traditional automation scripts become brittle over time. EvoTest employs a cognitive test automation framework powered by AI. This framework can not only self-heal from minor UI changes but also learn from successful test runs, continuously improving its efficiency.

 

4. Collaborative Testing Ecosystem:-

 

AI-powered Testing Hub with Gamification: EvoTest fosters collaboration through a central AI-powered hub. This hub provides:

Interactive Insights Dashboard: Real-time test execution reports with actionable insights based on AI analysis and historical data, visualized in a user-friendly format.

Conversational Testing Assistant with Gamification: A built-in conversational AI assistant, "Test Sensei," guides testers, resolves issues, and offers suggestions based on historical data and best practices. Gamification elements like leaderboards and badges promote healthy competition and knowledge sharing.

 

5. Continuous Learning and Improvement:-

 

Community-driven Knowledge Base: EvoTest leverages a community-driven knowledge base. Testers can contribute their learnings and insights, constantly enriching the system's knowledge and improving its future predictions and recommendations.

Regular AI Model Updates: EvoTest's AI models are continuously updated and improved based on new data, user feedback, and advancements in AI research. This ensures the solution stays at the forefront of ERP testing technology.

 

Proof-of-Concept: Explainable Prioritization

 

Imagine testing a new ERP module for inventory management. EvoTest analyzes code and historical data, predicting a potential defect in stock level calculations. The "Why" explanation highlights a specific mathematical formula within the code. Knowing the root cause, developers can prioritize fixing the formula, ensuring accurate inventory management before deployment.

 

EvoTest: Your Advantage

 

By leveraging the power of evolving AI, EvoTest empowers you to:

 

Outpace Testing Challenges: Continuously learn and adapt to evolving testing needs.

Reduce Risks Proactively: Predict defects with explainability, enabling targeted fixes.

Boost Collaboration and Knowledge Sharing: Foster a collaborative testing environment.

Optimize Efficiency and Test Coverage: Automate tasks and generate highly efficient test data and scripts.

 

EvoTest is your key to achieving successful ERP implementations through a constantly evolving and adaptive AI-powered testing solution. 

Userlevel 1

Strata: A Multi-Layered AI Approach

 

ERP testing can be a labyrinth of complexities. Strata cuts through the maze with a multi-layered AI solution that streamlines processes and empowers informed decision-making. Here's how Strata tackles the five key challenges:

 

1.Context-Aware Test Data Generation

Domain-Specific Knowledge Integration:- Strata goes beyond simple data generation. It leverages Natural Language Processing (NLP) to analyze user roles, business processes, and domain-specific terminology within the ERP system. Imagine simulating a purchase order with realistic product details, pricing, and approval workflows specific to your industry.

Adaptive Mutation Engine:- Strata employs a cutting-edge mutation engine powered by advanced AI. This engine injects realistic anomalies and unexpected user behavior into test data, mimicking potential disruptions like network issues or hardware failures. This ensures comprehensive testing for even the most unexpected scenarios.

 

2.Multi-Faceted Defect Prediction with Confidence Scoring

Heterogeneous Data Fusion:- Strata doesn't rely solely on historical testing data. It utilizes a sophisticated AI model that merges diverse data sources including code analysis, user feedback, and real-time system logs. This holistic approach provides a more comprehensive picture of potential defects.

Confidence-Scored Defect Prioritization:- Strata goes beyond simple defect prediction. It assigns a confidence score to each prediction, indicating the likelihood of the defect being real. This empowers testers to prioritize critical issues with the highest confidence levels for maximum impact.

 

3. AI-powered Test Automation with Self-Healing Capabilities

NLP-driven Test Case Generation with Explainability:- Strata utilizes NLP to analyze user stories, business requirements, and existing test cases. It not only generates individual scripts but also offers explanations for the generated steps. This allows testers to understand the reasoning behind the tests and adapt them if needed.

Cognitive Test Automation with Self-Healing:- Strata's test automation framework utilizes AI for self-healing capabilities. It can automatically adjust to minor UI changes in the ERP system, ensuring the continued effectiveness of your automated tests without the need for constant manual updates.

 

4. Collaborative Testing Hub with Knowledge Sharing

Interactive Insights Dashboard with Actionable Recommendations:- Strata's central AI-powered hub provides real-time test execution reports visualized in an interactive format. It goes beyond basic data by offering actionable insights based on AI analysis and historical data.

Conversational Testing Assistant with Community Forum:- Strata fosters collaboration through a built-in conversational AI assistant, "Test Navigator." Test Navigator can guide testers, answer questions, and suggest solutions based on historical data and best practices. Additionally, a community forum allows testers to share knowledge and best practices, fostering continuous learning.

 

5. Continuous Learning and Improvement

 Federated Learning for Decentralized Knowledge Sharing:- Strata utilizes a federated learning approach. This allows secure knowledge sharing between different deployments of Strata without compromising sensitive data. This shared knowledge pool continuously improves the AI models deployed across various organizations.

Regular AI Model Updates with Explainable Enhancements:- Strata's AI models are continuously updated and improved based on new data, user feedback, and advancements in AI research. Additionally, new enhancements are explained in an understandable way, empowering users to trust the AI's decision-making.

 

Proof-of-Concept: Confidence-Scored Prioritization

Imagine testing a new ERP module for financial transactions. Strata analyzes code and historical data, predicting a potential defect in currency conversion calculations. However, the confidence score assigned is relatively low. This allows testers to prioritize a different defect with a high confidence score related to user authentication, ensuring critical issues are addressed first.

 

Strata: Your Competitive Edge

By leveraging a multi-layered AI approach, Strata empowers you to:

Achieve Superior Test Coverage:- Context-aware data generation and adaptive mutation engines ensure comprehensive testing of realistic scenarios.

Focus on High-Impact Issues:- Confidence-scored defect prediction allows for prioritizing critical defects with the highest likelihood of causing problems.

Boost Collaboration and Knowledge Sharing:- The collaborative testing hub fosters a culture of learning and continuous improvement.

Reduce Manual Work and Optimize Efficiency:- Automate tasks like test data generation and case creation, freeing up resources for more strategic activities.

 

With Strata, gain a competitive edge in your ERP testing and ensure a smooth and successful implementation. 

Imagine a Sherpa for your ERP testing journey. Not just a guide, but an AI-powered assistant that anticipates challenges, optimizes your path, and equips you to conquer the peak of successful implementation.

 Sherpa - The ERP Solution:

  • Test Data Management:
    • Leverage AI to analyze historical data and industry trends,¬†predicting potential data anomalies and security concerns.
    • Utilize natural language processing (NLP) to understand your specific business needs and generate tailored test data sets that reflect real-world scenarios.
  • Predictive Test Case Prioritization:
    • Train Machine Learning (ML) models on historical testing data to identify areas with the highest risk of defects.
    • Integrate with business intelligence tools to understand the criticality of different functionalities to your operations.
    • Dynamically prioritize test cases based on risk scores,¬†business impact,¬†and real-time test results,¬†ensuring you address the most critical issues first.
  • Intelligent Test Automation:
    • Implement AI-powered tools that can learn from existing test scripts and user actions,¬†automating repetitive tasks and creating new test cases on the fly.
    • Utilize self-healing algorithms to identify and fix minor errors within automated scripts,¬†minimizing downtime and ensuring smooth test execution.
  • Collaboration and Insights:
    • Foster collaboration through a centralized platform where testers,¬†developers,¬†and business stakeholders can access real-time testing data and insights.
    • Employ NLP-powered chatbots to answer tester queries,¬†provide contextual help,¬†and facilitate knowledge sharing within the team.
    • Leverage AI to generate automated reports with visualizations highlighting potential defects,¬†root cause analysis suggestions,¬†and recommendations for improvement.

Benefits of Sherpa : 

  • Predictive Maintenance:¬†

AI analyzes testing data and user behavior patterns to identify potential system issues before they occur. This allows for proactive maintenance and prevents disruptions during ERP implementation.

  • Continuous Learning:¬†

The ERP Sherpa continuously learns and evolves based on testing results and user feedback. This ensures the system adapts to changing business needs and testing methodologies.

  • Reduced Risk and Costs:¬†

Proactive risk identification and intelligent test prioritization minimize the chance of critical defects slipping through, leading to cost savings and a smoother implementation.

  • Improved Quality and Agility:¬†

AI-powered data generation, automation, and insights ensure comprehensive testing, enabling faster feedback loops and a more agile implementation process.

  • Enhanced Collaboration and Transparency:¬†

The centralized platform and AI-powered communication tools break down silos and foster collaboration between all stakeholders.

  • Centralized Platform:¬†

Sherpa  provides a centralized platform where testers, developers, and business stakeholders can collaborate seamlessly. This platform serves as a hub for sharing information, discussing issues, and tracking progress, fostering a cohesive team environment.

  • Real-Time Data Access:¬†

With Sherpa , stakeholders have access to real-time testing data and insights. This ensures that everyone is working with the most up-to-date information, facilitating informed decision-making and coordinated action.

  • Automated Reporting:

 Sherpa  generates automated reports with visualizations highlighting potential defects, root cause analysis suggestions, and recommendations for improvement. These reports serve as a common reference point for all stakeholders, enabling them to align their efforts towards common goals.

  • NLP-Powered Chatbots:

 Sherpa  employs NLP-powered chatbots to provide contextual help and facilitate knowledge sharing within the team. These chatbots can answer questions, provide guidance on testing processes, and offer assistance whenever needed, enhancing communication and collaboration.

  • Conversational Defect Tracking:¬†

Sherpa  facilitates conversational defect tracking, allowing stakeholders to discuss and resolve issues in real-time. This collaborative approach ensures that defects are addressed promptly and effectively, preventing them from escalating into larger problems later on.

 

The ERP Sherpa isn't just a tool, it's your trusted companion on the path to successful ERP implementation.

 

Userlevel 1

Excited to see the submissions, Thanks to the QA community for participation

We are excited to present this month challenge created by @sidharth shukla on how AI can help with challenges in the ERP testing. 

Enterprise Resource Planning (ERP) systems are complex and extensive, making testing a formidable task. The intricacy of these systems presents unique challenges such as test data management, test case prioritisation, and test automation. Your challenge is to devise an Artificial Intelligence (AI)-based solution to address these complexities, thereby enhancing the efficiency and effectiveness of ERP testing.

Background Information:

  1. ERP systems: ERP systems are comprehensive software solutions that integrate and automate diverse business operations, including finance, human resources, supply chain management, and customer relationship management.
  2. ERP testing challenges: The complexity of ERP systems leads to issues such as managing voluminous and varied test data, prioritising test cases for maximum coverage, and automating tests in a resource-efficient manner.
  3. AI in ERP testing: AI technologies like machine learning, natural language processing, and computer vision can revolutionise ERP testing by automating test data management, predicting defects, optimising test coverage, and streamlining test automation.

Challenge Objectives:

  1. Suggest an AI-based solution that addresses the key challenges in ERP testing, including test data management, test case prioritisation, and test automation.
  2. Define an approach that leverages AI to generate relevant test data, predict defects, and prioritise test cases based on their impact on business processes and risk of failure.
  3. Demonstrate how your AI solution improves collaboration between various stakeholders involved in the ERP testing process, thereby enhancing the overall quality and agility of ERP implementation.

Prizes and Points:

ūüŹÜ 2 Winners: Personalized Certificate of Achievement signed by Sidharth, +300 points, a badge, and a gift box from us.

ūüĆü All Participants: +150 points for your valuable contribution to the challenge.

Submission Requirements:

  1. No need to design and develop any solution using AI, rather pick the problem statement and try to come up with a write up on How AI can be used to solve or minimise the efforts. We are specially looking for five  important points: test data management, test case prioritisation, and test automation, predict defects, collaboration between various stakeholders
  2. A proof-of-concept implementation showcasing your solution's key functionalities, such as test data generation, defect prediction, or test case prioritisation.

Evaluation Criteria:

  1. Innovation: The originality and creativity of your approach in leveraging AI to address ERP testing challenges.
  2. Practicality: The feasibility of implementing your solution in real-world ERP testing scenarios, considering constraints like cost, time, and resources.
  3. Impact: The potential of your solution to significantly improve the efficiency and effectiveness of ERP testing and its applicability across various industries and ERP systems.

Key Dates:

  • Challenge release: May 8
  • Lasts: 3¬†weeks
  • Judging (for ShiftSync members): 3 days
  • Winners announced: June 4

Example

AI can analyse historical data and generate synthetic data that accurately represents real-life scenarios, reducing the time and effort required for test data management in ERP systems.
Imagine a company that uses an Enterprise Resource Planning (ERP) system to manage its business operations. As part of their testing process, they need to generate large amounts of test data to simulate various scenarios, such as inventory management, sales transactions, and financial reporting.

Traditionally, creating this test data manually can be time-consuming and labor-intensive. However, by leveraging AI technology, the company can streamline this process significantly.

Using AI algorithms, the system can analyze historical data from the ERP system and identify patterns, trends, and correlations. It can then use this information to generate synthetic data that accurately represents real-life scenarios.

For example, if the company wants to simulate a busy sales day with high transaction volumes, the AI can generate synthetic sales orders, customer data, and inventory records based on past sales data. This synthetic data closely mimics the characteristics of real data, allowing the company to perform thorough testing without the need to manually create or manipulate test data.

By automating the generation of test data using AI, the company can save time and effort, improve the efficiency of their testing process, and ensure more comprehensive test coverage in their ERP system.

Disclaimer: The use of generative AI to solve this challenge, is allowed in the context of formulation but NOT ideation. Every answer should have a clear "human element" otherwise said answer will be disqualified. 

Let the Challenge Begin!

Enterprise Resource Planning (ERP) systems are vital for integrating and automating various business processes, including finance, human resources, supply chain management, and customer relationship management. Given the complexity and extensive nature of ERP systems, testing these systems is a formidable task. Key challenges include managing vast amounts of test data, prioritizing test cases, automating tests, predicting defects, and enhancing collaboration among stakeholders. This document outlines an AI-based solution to address these challenges, thereby improving the efficiency and effectiveness of ERP testing.

1. Test Data Management
Challenge:
Managing voluminous and varied test data is crucial for effective ERP testing. Traditional methods of generating and managing test data are time-consuming and prone to errors.

AI Solution:
AI-Driven Synthetic Data Generation
AI algorithms can analyze historical data from ERP systems to identify patterns and correlations. Using this information, AI can generate synthetic data that accurately represents real-life scenarios.

Implementation Example:

Data Pattern Analysis: AI models analyze historical transaction data to understand patterns such as peak sales periods, inventory turnover rates, and financial transaction trends.
Synthetic Data Generation: Based on the analysis, AI generates synthetic data that mimics real-world data. For example, simulating a high-transaction volume day by creating synthetic sales orders, customer data, and inventory records.
Benefits:

Reduces time and effort in manual data creation.
Ensures comprehensive and realistic test coverage.
Enhances the accuracy of test results by using data that closely resembles actual business scenarios.
2. Test Case Prioritisation
Challenge:
Prioritizing test cases to ensure maximum coverage and risk mitigation is complex due to the extensive functionality of ERP systems.

AI Solution:
AI-Based Test Case Prioritisation
Machine learning algorithms can prioritize test cases based on historical defect data, usage patterns, and critical business processes.

Implementation Example:

Risk Assessment: AI evaluates the risk associated with different modules based on past defect data and business impact.
Usage Pattern Analysis: Analyzes usage data to identify frequently used functionalities that are critical for business operations.
Prioritisation: AI ranks test cases by considering risk assessment and usage patterns, focusing on high-impact areas first.
Benefits:

Ensures critical functionalities are tested thoroughly.
Optimizes testing efforts by focusing on high-risk areas.
Reduces the likelihood of critical defects in production.
3. Test Automation
Challenge:
Automating tests in a resource-efficient manner while ensuring comprehensive coverage is challenging due to the dynamic nature of ERP systems.

AI Solution:
AI-Powered Test Automation Framework
AI can enhance test automation frameworks by optimizing test scripts and maintaining them with minimal human intervention.

Implementation Example:

Dynamic Test Script Generation: AI generates test scripts based on functional requirements and user stories.
Self-Healing Scripts: AI models detect changes in the ERP system and automatically update test scripts to accommodate these changes.
Test Optimization: AI identifies redundant test cases and optimizes the test suite to improve efficiency.
Benefits:

Reduces manual effort in maintaining test scripts.
Ensures continuous and efficient test automation.
Enhances test coverage and accuracy.
4. Predicting Defects
Challenge:
Predicting potential defects early in the testing cycle is crucial to prevent costly fixes in production.

AI Solution:
AI-Driven Defect Prediction
Machine learning models can predict potential defects by analyzing historical defect data, code changes, and development practices.

Implementation Example:

Defect Pattern Analysis: AI analyzes historical defect data to identify common defect patterns and root causes.
Predictive Modeling: Uses data from current development and testing cycles to predict areas likely to have defects.
Preventive Measures: Provides insights and recommendations to prevent predicted defects.
Benefits:

Identifies potential defects early, reducing the cost of fixes.
Improves overall software quality.
Enhances the efficiency of the testing process by focusing on high-risk areas.
5. Collaboration Between Stakeholders
Challenge:
Ensuring effective collaboration between various stakeholders involved in ERP testing is essential for a successful implementation.

AI Solution:
AI-Enhanced Collaboration Platforms
AI can enhance collaboration by providing insights, automating communication, and facilitating real-time updates and feedback.

Implementation Example:

Insight Generation: AI analyzes test results and generates actionable insights for developers, testers, and business analysts.
Automated Communication: AI-powered bots automate communication, providing updates on testing progress, defect status, and test results.
Real-Time Collaboration: AI facilitates real-time collaboration by integrating with collaboration tools like Slack, Jira, and Microsoft Teams.
Benefits:

Enhances transparency and communication among stakeholders.
Improves decision-making with data-driven insights.
Accelerates the testing and development process by ensuring timely updates and feedback.

 

Proof-of-Concept Implementation
To showcase the key functionalities of our AI-based solution for ERP testing, we will outline a high-level proof-of-concept (PoC) implementation. This PoC will focus on three main areas: test data generation, defect prediction, and test case prioritization.

1. Test Data Generation
Objective:
Automate the generation of realistic and varied test data to ensure comprehensive test coverage.

Implementation Steps:

Data Collection:

Collect historical data from the ERP system, including transaction records, user activity logs, and business process data.
Pattern Analysis:

Use a machine learning model to analyze the collected data for patterns, trends, and correlations. Techniques such as clustering and association rule mining can be used.
Synthetic Data Generation:

Develop an AI model (e.g., Generative Adversarial Network, GAN) to generate synthetic test data that mirrors the identified patterns and trends.
Implement a script to generate data scenarios like high transaction volumes, month-end financial reporting, and peak sales periods.
Tools and Technologies:

Python, TensorFlow, or PyTorch for machine learning models.
Pandas and NumPy for data manipulation.
SQL/NoSQL databases for data storage and retrieval.

 

 

Badge

ERP Testing AI Model: Defect Prediction & Priority Based Verification

Predicting defects and prioritise verification based on the impact on business processes are important for an effective ERP testing AI model.

 

Historical Data analysis: 

Historical data holds up many flaws which can help to create an effective test strategy. So, its always beneficial to explore historic data.

There are many tools available online paid & open source for doing ‚ÄúHistorical data analysis‚ÄĚ. Since ERP applications are always closer to an organisation data and one needs to be cautious while sharing those information to any tool. So, I always advocate of using anything built in-house.¬†

 

AI solution (Wish List): 

1. Data Collection: Gather historical data from relevant resources like: Defects, logs, Test Results

2. Data Analysis: Understand the pattern from gathered data

This step will completely depend on 

 - How well the defect template is/was defined in Jira or any other defect tracking tool.

 - Are we tracking defect RCA, environment details and Defect classification (Functional, Performance and Security)

3. Model (Logic): 

- Algorithm to predict defect occurrence. 

- Train the model to learn pattern from past data to make future prediction.

- Test the model with new data to predict the occurrence of defects in newly developed features.

- Continuously monitor the model to check the accuracy.

 

Prioritise Test Verification 

Test prioritization based on the impact on business processes can be a complex task which involves both manual and automated steps.  

The initial identification and mapping of business process can be manual and which can be used later to train the AI models.

 

1. Business Process - Identification

a. List out all critical business process within the ERP system.

b. List the dependency/relationship between business processes and modules.

 

2. Business Process - Impact Analysis

a. For the above identified processes assess the business impact which can include factors like frequency of use, user impact and revenue impact.

b. Use the above data collection to understand which process or module have the most impact when failure occurred.

 

3. Testcase Identification:

a. Identify the test cases for the above identified business process and list out execution frequency and past failures.

 

4. Model(Logic)

 - Extend the algorithm to predict the occurrence of defect in identified processes based on the historic data.

 - Based on the defect likelihood develop a scoring algorithm to prioritise the business process related test cases.

 - Focus testing on high priority test cases which has significant business impact.

 - Continuously monitor and retrain the model.

 

Team Collaboration

A god tool or solution is useless if it is not used correctly. 

For a well defined AI model, collaboration among different team members plays an important role. Here’s how each team member can play their role well.

 

Business Stakeholders/PO:

They possess valuable domain expertise. Business stakeholders can review AI model outputs to make sure model align with business goals. 

Business stakeholders/PO’s can execute scenarios under various business conditions.

 

QA:

Provide feedback on model by performing rigorous testing. Identify areas of improvement.

 

PM:

Being a facilitator they can initiate communication among stakeholders and ensure alignment with project milestone.

Reply