Synechron bets large on AI in testing

The New York software consultancy with financial services focus is investing in predictive analytics and natural language processing to differentiate itself in an increasingly commodified market, says Ashish Nangla (pictured), senior director at Synechron.

Synechron, a New York-based IT consulting company focused on the financial services, is investing in artificial intelligence technologies in testing. The technology firm hopes to use the technology to power predictive analytics and natural language processing to achieve savings of 30% for clients of their testing services, said Ashish Nangla, senior director at Synechron.

While the company already has a few of projects that use AI technologies in testing, the technology will be rolled out the entire its quality assurance division this summer, between June and July. Testers will be provided with the technology and training to implement machine learning in the projects that they are working on.

The drive for AI, said Nangla, was spurred by Synechron’s need to differentiate itself in an increasingly commodified market: “It’s hard to stick-out in quality assurance nowadays. That is why AI has a good use-case: it can be a differentiator. So far the only people doing AI in testing are the big players or very niche firms. We’re ideally positioned to use it to lower the price of our services to give ourselves a competitive edge. This is our big gamble”

The first application of AI at Synechron was rolled out with a client in the insurance space which the firm had been working with for six years. The large amount of historical test data made the project an ideal candidate for applications of AI said Nangla. In this case, the technology was used to power predictive analytics.

“We took the data and created a machine learning model that looked at software failures and their root-causes,” said Nangla. The end result is a system that integrates with Jira, a quality management tool. The machine learning system is able to identify problem-areas and score proposed new features in the software delivery pipeline based on probability of failure.

According to Nangla the test team achieved a reduction in man-hours necessary, concentrating testing resources on those section of code that are most likely to break: “In two development cycles we have been able to reduce the amount of testing we conduct by 30%. And as we gather more data to power the machine learning algorithm, we can achieve even greater efficiencies,” he said.

It was the success of this project that spurred Synechron’s investment into AI for testing, which now includes a natural language test case processing. The idea is to use AI to translate business analyst requirements from natural language (such as English) into an executable test case. The quality assurance specialist who would normally translate the specifications laid out by business analysts into a testing schedule, is only necessary to verify the document.

While there is currently a team working full-time to research and develop new applications of AI in the software development lifecycle, enterprise-wide roll-out of AI will take another year to year and a half, said Nangla.

Tweet about this on TwitterEmail this to someoneShare on LinkedIn

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>