How AI-driven Quality Engineering is Driving Value Across the Software Development Life Cycle
Artificial Intelligence (AI) and Machine Learning (ML) are becoming increasingly pervasive thanks to organizations across industries rapidly adopting applications built on these emerging technologies. Businesses are already seeing significant value from implementing these technologies. Let’s deep dive into a few use cases across several industries that Mindtree currently serves to understand their common characteristics and analyze them from the lens of quality engineering.
The travel and hospitality domain has been one of the early adopters of AI/ML techniques. In its quest to provide greater business value to its customers, the industry uses AI/ML to enhance customer experience by providing highly personalized recommendations around optimal pricing for flights, hotels, rental cars and holidays - based on their buying patterns, behavior and preferences. Data sets (see Figure 1 for data sources) that AI/ML algorithms leverage are highly diverse and complex, making such personalized recommendations humanly impossible.
Figure 1: Varied sources of data for powering AI/ML algorithms
One the biggest challenges confronting insurance companies is pricing the policies correctly. For example, for auto insurance, policy rates are typically determined by simple factors such as vehicle, model, and year of make. This method doesn’t factor in individual behaviors such as driving habits, location, and weather which can significantly impact the accuracy of the rates. Getting these data points and analyzing them manually or through traditional tools is not only cumbersome but also error prone. Al algorithms can parse complex data, find anomalies, and extract patterns of driving behavior to build recommendations for individualized policy rates. AI can also be used to evaluate a variety of factors for assessing the risk of an accident.
Media and Entertainment
Organizations are training ML algorithms for developing film trailers and designing advertisements. Yet another application of AI in this industry is recommending personalized content based on user activities. Media content providers are also using AI software to improve the speed and efficiency of media production processes and the ability to organize visual assets.
Retail businesses are looking to make highly personalized product recommendations for various types of customers based on customer data collected from various sources such as website homepages, searches, kiosks and shopping malls. AI helps create market segments across types of customers, patterns in customer’s buying behavior from previous purchases, demographics and preferences. The segments are then used for developing personalized recommendation that are delivered through customer accounts on websites, email campaigns, call center agents and location-based advertisements across multiple channels (see Figure 2).
Figure 2: Customized product recommendations - sample
While these are distinct use cases that target specific industry customers, we can further analyze the common characteristics across these cases for quality engineering across domains. Some of these traits include:
- Extensive customer data stored in multiple repositories can be analyzed adequately to provide useful insights in terms of customers’ preferences, buying patterns and behaviors. These insights then become inputs for application development aimed at delivering greater business value that customers are looking for.
- All of these domains have a complex mesh of applications, devices, systems and interfaces – that work in tandem to deliver business value to the customers. This intertwined mesh of “things” adds to the complexity of application development, delivery, deployment and support.
- As systems become more complex and customers more demanding, the pressure on software delivery becomes even greater. This has led to the creation of nimble architecture and adoption of Agile/DevOps on the one hand. On the other, it has led to the accumulation of undeveloped product features and defects backlog, adding to technical debt. This has a bearing on how applications are tested. AI can help determine the priority and hot spots for testing.
In essence, AI-driven quality engineering delivers value to the software development life cycle (while these domains are re-engineered by AI) by enabling the following capabilities:
- Forecasting customer’s requirements and dynamically tuning the test strategy.
- Scanning the use cases and test case repository for inclusion of relevant test scenarios.
- Driving extreme automation – identifying scenarios from UI, Services, and API levels to maximize automation.
- Developing defect prediction model based on criticality and previous downtimes of applications.
- Validating features for performance, security and stability.
To know more about how Mindtree experts can help you, click here.