top of page

Art of Intelligent Automation

Reshaping the future of work with robots

As Business Applications become complex, so the testing going to be. With traditional test practices its next to impossible to maintain quality of business applications in long run.

Keeping same in mind, now Companies like to introduce automation as soon as development gets started.

This approach not only help in maintaining better quality of business applications but also result in significant cost savings in terms of time and effort.

And there is no better solution other than Intelligent Automation.

This article we will share some be best practices which Next Generation Automation team proposed to the clients and done some successful Proof of Concept in recent past.

Practice 1: Structured Data Interaction (SDI)

These are traditional systems where the integration is through exchange of information that is well structured. Examples include integration of systems through relational data base management systems ( RDBMS) , data transformation tools, and application programming

interfaces ( APIs) and web services.

More intelligently we streamline structured data interaction using automation among both homogeneous and heterogeneous systems, better the flow of information and resulting better product quality.

Practice 2: Robotic Process Automation

It involves automation of standardized and rules driven system-based activities using scripts and other methods to support efficient business processes. It is suitable in scenarios where it is too expensive or inefficient for humans to execute a task or a process. Very helpful for tasks that involve repetition with out change.

Use Case for Testing can be All failures tests automatically converted into test defects and submitted to Defect tracking tools like Jira for defect analysis and defect fix.

Practice 3: Machine Learning

It involves systems that learn through handling variations that are not anticipated upfront. These systems get trained on the go by assimilating leanings from the data and decisions, and may make simple predictions or classifications backed by algorithms. A simple case could be a scenario where a well-defined identifier needs to be mapped to more descriptive/free form text, e.g., mapping of a vendor name that appears on an invoice to the vendor ID in the system. The vendor name may appear in various forms.

Machine Learning results significance reduction in test effort if implemented rightfully in your application test cycles.

Practice 4: Natural Language Processing

NLP uses statistical methods and learning algorithms to analyze text and unstructured information to understand the meaning, sentiment and intent. A sample use case could be the customer service function, where a customer raises a support ticket in form of free text, which is analyzed to understand and determine the levels of urgency, sentiment or frustration and then determine the ticket severity/priority.

Other scenario can be understanding system requirements and generate test cases. Or check whether system requirements written correctly as per customer Request for Information (RFI) document.

Practice 5: Natural Language Generation

It is a technology that helps generate text as we speak or write from structured information such as fields and numerals. It is largely applied where sections of financial analysis reports and insights are generated, e.g., numbers reflecting a company’s performance. During Business meetings it can be applied to convert client expectations into meaningful business requirements.

Practice 6: Chat bots and virtual agents

These are systems that can interpret voice/text in free form (chat) to simply respond with standard predefined answers. A simple example is the customer service function where a chatbot could respond to queries. These chat bots can continuously learn and build vocabulary to interpret unstructured information being directed to them.

For Testing use case can be, Product owner write to chat bot show me last 5 release test results or show me defect density for last release or show me average execution time per test script or show me number of defects generated from automation scripts.

Practice 7: AI Decision Systems

These are systems that employ an array of technologies, algorithms and models to solve complex and inter-related problems to make decisions. These may be driven by deep learning systems and cognitive capabilities to recognize patterns, and apply statistical models and algorithms to make choices and decisions. These could also potentially address multiple decision points, e.g., determining the demand for certain products for a geography/location based on weather forecasts, thereby helping decide the inventory to be housed in a store and determine the best possible fulfillment center location and route to be chosen for the fulfillment.

Use case for Testing can be Building auto generated Object repository and apply AI algorithms to determine whether your application holds valid automation ids

that going to help automate your application. Many scenarios developers forget to add unique ids to all application components in order to save development time but it makes automation fragile. Such scenarios can be looked at if AI Decision systems in place at very start of Development Life Cycle.


Building better QA for tomorrow


bottom of page