Accenture plans to infuse greater amounts of artificial intelligence into application development testing following its acquisition of Real Time Analytics Platform Inc., which provides tools that make use of machine learning algorithms, neural networks and natural language processing (NLP) to analyze every stage of the software development life cycle.
Kishore Durg, senior managing director and growth and strategy lead for Accenture Technology, the IT services arm of Accenture, will incorporate the AI technologies developed by Real Rime Analytics Platform into the Accenture Touchless Testing Platform, which is a suite of open source and proprietary testing tools that Accenture has infused with advanced analytics and cognitive computing technologies.
Durg said Accenture is committed to automating application testing using AI technologies to help organizations increase the productivity of application development teams. At a time when organizations are investing in a broad range of digital transformation projects, the rate at which applications can be built, tested and deployed has become a critical issue.
Most organizations on their own won’t be able to aggregate enough data to drive machine learning algorithms and neural networks to drive a meaningful AI model. Those models require access to massive amounts of data to be trained to identify potential issues within any application. Because of the size and scope of the projects Accenture works on, organizations that employ its service can benefit from all the data that Accenture can bring to bear to train AI models, Durg noted.
The quality of the data used to train those AI models is also as important as the amount of data, added Durg. Accenture has the resources required to cleanse data before it gets fed into an AI model, as well the expertise needed to eliminate any biases that might get built into any algorithm being applied to that data, he said.
As AI continues to evolve, it won’t be long before every application incorporates some form of an AI model. Each of those models not only needs to be trained, but they also need to be updated continually—and eventually replaced by more sophisticated models as more data becomes available. In effect, that requirement will create a need to extend DevOps processes to how AI models are incorporated within an application.
At the same time, it’s clear that large swaths of DevOps processes involving, for example, application testing can be increasingly automated using a variety of machine learning and network networks, also known as deep learning algorithms.
In theory, that should accelerate the rate at which applications are developed and deployed, while testing processes that are augmented by AI technologies should result in applications that are more secure, as many more vulnerabilities will be discovered long before an application gets deployed in a production environment.
In fact, the inclusion of advanced algorithms across a DevOps toolchain might accelerate adoption of DevOps processes. After all, most IT organizations today have embraced DevOps unevenly largely because of the complexity involved. AI technologies one day soon may eliminate that issue altogether.
