At this very moment, the software of autonomous testing company AutonomIQ is scouring the internet to learn about every application on the network, building a natural language dictionary to understand every software it tests.
We coined the term ‘autonomous testing’
This immense data gathering is just one cog on the wheel of the software life cycle from developing to testing, releasing, operating, and back round again.
The Sociable caught up with the Co-Founder and CEO of the company that coined the phrase “autonomous testing,” Ram Shanmugam, at the Horasis India Meeting in Malaga, Spain to discover how AI and Machine Learning are disrupting the life cycle of software.
From Left to Right: Vineet Mohan, Chairman at Armstrong; Balvinder Singh Powar, Professor at IE Business School; Srikar Reddy, CEO at Sonata Software; and Ram Shanmugam, CEO and Co-Founder at AutonomIQ
“AutonomIQ is using artificial intelligence and machine learning to disrupt the software testing life cycle. So, if you think about software testing, it’s a $60-$70 billion market, and 80 to 90 percent of that is done by manual, human testing,” says Shanmugam.
We are helping companies achieve release cycles by shrinking the time it takes to test
This is a large chunk of time dedicated to labor that traditionally, at least in the fast-paced digital sense, has been done by humans.
“Software testing was one of the spaces where we found huge amounts of manual work that could be automated,” the AutonomIQ Co-Founder said.
And automate they did!
AutonomIQ came up with ways to create test cases, data, and scripts in a matter of minutes instead of hours or days as had been done with traditional models.
At least that’s what it says on the website, but if you ask Mr. Shanmugam, those cycle times will be greatly reduced very soon.
“In the next six to 12 months, we’ll be able to do it in a few minutes,” he says.
Our five year view of technology is that machines should be able to code
This is the speed at which AutonomIQ’s team is progressing using AI and Machine Learning to test software, going from cycle time reduction of 200 hours to just five hours as it is presently, to being able to do it in a matter of minutes in the coming year. And there are already plans for the next phases beyond testing.
“We are helping companies achieve release cycles by shrinking the time it takes to test. Beyond testing, our next intent will then be to go on into the release management function, which is when the product is tested and then gets deployed. We want to disrupt the release management part of the life cycle. And eventually we want to go after the operations side.”
If you think about the last industrial revolution, there were two pockets — the design element of the car, and then there was the production element of the car. We see software like that.
The AutonomIQ CEO explained that the operations side occurs after software has been running in the data center or in the cloud.
“If something breaks, can it be detected and can it be brought back into the development life cycle?”
In other words, AutonomIQ is working towards a circular model of software testing, release management, operations, and then back into the development phase where the software will need to be tested again, and the cycle continues.
Our five year view of technology is that machines should be able to code as well
“If you think about the last industrial revolution, there were two pockets — the design element of the car, and then there was the production element of the car. We see software like that. There is the design element of software that is coding, and then there is the industrialization of software, which is taking the code, testing it, releasing it, and managing it once it’s released,” Shanmugam explains.
“So that industrialization process of software is today done by human beings, and it is very manual, and it takes a long time, and costs a lot of money, but we are essentially trying to use machines for the entire industrialization process.”
“Our five year view of technology is that machines should be able to code as well. I can write the requirements, and from the requirements it will detect what are all the things I will need to build, and it starts developing it, and it starts executing it, and then the whole life cycle starts.”
