IBM’s Full AI Cycle Testing solution is a combination of unique set of methods and processes supported by a set of Artificial Intelligence infused tools to support the testing of AI Infused Applications. This approach focusses on testing every building block of an AI based application and ensure quality assurance of the same.

• Data quality assessment through metrics based standardization
• Generation of optimal test cases to achieve maximum test coverage.
• AI/ML Model assessment beyond the accuracy assessment to validate the fairness and robustness of the models.
• End to End test automation using industry standard automation tools and unique AI based tools.
• Defect and Test Result Analysis brings in more insights to the outcome of the testing.

An extensive list of AI based tools and assets that cover end to end procedure are being used. The capabilities included in IBM’s Full AI-Cycle Testing is listed below:

• Requirements Analytics focusses on predicting the complexity of requirements to create the test plan strategy based on the same.
• Data Quality Assessment evaluates the quality of structured or unstructured data through metrics-based approach.
• Test Model Enhancer recommends the structured representation of unstructured input and enables in creating equally dense data combinations for testing.
• AI Model Evaluator is used to assess the fairness of AI/ML model through sensitivity and adversarial robustness evaluation.
• Defect Analytics provides a real-time defect root cause predict and defect pattern analytics to provide recommendations to decrease defect density.
• Test Results Analysis will provide insights on identifying underperforming test scenarios and root causes for the same
• Payload monitoring capability to evaluate dialogue path coverage, incomplete path analysis and test case generation using payload logs for conversation services.

Value Proposition:

• A complete set of AI and non-AI based testing tools that are focussed on enhancing testing team’s productivity and effectiveness
• These tools cover end to end quality engineering capabilities across all phases of testing for AI Infused Application.
• All these tools are hosted on a Quality Platform and provided as cloud services or on-premise service
• Foundational technical skills are sufficient to use these tools
• A wide range of AI Based Applications are supported by these set of tools
  • Conversation Based solutions
  • Speech and Audio based AI solutions
  • Content Digitization
  • Images based AI solutions including Edge applications
  • NLP and NLC based applications.
  • Intelligent Workflow based applications
  • Human in the loop based solutions

Ilan Jayaraman has 19 plus years of experience on wide range technologies from Mainframes to Artificial Intelligence. In the span of his career, he has performed multiple roles ranging from developer to thought leader. He is currently the Practice Lead for AI Testing at IBM. He started his journey as a Mainframe Developer and progressed his career performing multiple roles as Project Manager, DPE and Product Manager before taking over as Practice Lead for AI Testing.

In the role of Product Manager, he has lead the development of IGNITE Quality Platform. This platform is an AI based set of tools to support and optimize the end to end testing processes at IBM. This platform supports Quality Engineer in delivering end to end testing methods and services to their client using a set of Deep learning, Clustering and NLP based AI tools.

As Practice Lead for AI Testing, he has come with the Point of View of testing AI Infused Applications. He has worked IBM Research and IBM Automation practice to develop Full AI-Cycle Testing approach to testing AI Infused Applications. In the scope of this role, he has been instrumental in collaborating with IBM Research to develop unique set of AI based assets to support this solution. In the capacity of this role, the process, methods and capability building strategy has been developed at IBM to support AI Infused Application testing.