To alleviate DevOps skills issues, we need more AI skills, ironically

Artificial intelligence is reportedly boosting intelligence within businesses and is also doing the same for information technology shops. For example, AIOps (artificial intelligence for IT operations) applies AI and machine learning to data streaming from IT processes, sifting through the noise to detect, spotlight, and head off problems. 

AI and machine learning are also finding a home in another emerging area of IT: assisting DevOps teams in assuring the viability and quality of the software that is moving at ever-faster speeds through the system and out to users. 

As found in a recent survey out of GitHub, development and ops teams are turning to AI in a big way to smooth the flow of code through the software review and testing phase, with 31% of teams actively using AI and ML algorithms for code review — more than double last year’s number. The survey also finds 37% of teams use AI/ML in software testing (up from 25%), and a further 20% plan to introduce it this year.

Also: Understanding Microsoft’s grand vision for building the next generation of apps

An additional survey out of Techstrong Research and Tricentis confirms this trend. The survey of 2,600 DevOps practitioners and leaders finds 90% are favorable about injecting more AI into the testing phase of DevOps flows, and see it as a way to resolve skills shortages they are facing as well. (Tricentis is a software testing vendor, with an obvious stake in the results. But the data is significant as it reflects a growing shift toward more autonomous DevOps approaches.)

There’s even a paradox that emerged from the Techstrong and Tricentis study: Enterprises need specialized skills in order to alleviate a need for specialized skills. At least 47% of respondents state that a major benefit of AI-infused DevOps is to reduce the skills gap, and “make it easier for employees to perform more complicated tasks.” 

Also: DevOps nirvana is still a distant goal for many, survey suggests

At the same time, a lack of the skills needed to develop and run AI-powered software testing was cited by the managers as one of the leading barriers to AI-infused DevOps, at 44%. This is a vicious cycle that hopefully will be remedied as more professionals participate in training and educational programs focused on AI and machine learning.  

Once AI does start getting put into place with IT sites, it will help make a dent in process-intensive DevOps workflows. Nearly two-thirds of managers in the survey (65%) say functional software testing is well suited to and would benefit greatly from AI-augmented DevOps. “DevOps success requires test automation at scale, which generates massive amounts of complex test data and requires frequent changes to test cases,” the survey’s authors point out. “This perfectly aligns with the capabilities of AI to identify patterns in large data sets and offer insights that can be used to improve and accelerate the testing process.”

Also: Artificial intelligence projects grew tenfold over the past year, survey says

Along with potentially reducing skills requirements, the survey also identified the following benefits to infusing more AI into DevOps:

• Improve customer experience: 48%
• Reduce costs: 45%
• Increase the efficiency of developer teams: 43%
• Increase code quality: 35%
• Diagnose problems: 25%
• Increase velocity of releases: 22%
• Codifying knowledge: 22%
• Prevent defects: 19% 

Early adopters of AI-augmented DevOps tend to be from larger organizations. This is not surprising, since larger concerns would have more developed DevOps teams and greater access to advanced solutions such as AI. 

Also: It’s time for technology teams to find their voice in customer experience

“In terms of DevOps, these mature companies are marked by the progress they’ve made in streamlining their software development capabilities over the past five to seven years and their mature and refined pipelines and processes,” the Techstrong and Tricentis authors point out. “These DevOps organizations are cloud-native and use DevOps workflow pipelines, toolchains, automation, and cloud technologies.”

In the long run, infusing AI to assist with vital aspects of DevOps is a smart idea. The DevOps process, for all its collaboration and automation, is only getting more exhausting as software is expected to fly out the door at a quickening pace. Leave it to the machines to handle a lot of the onerous aspects, such as testing and monitoring.