A global survey of 700 software engineering practices published this week finds that thanks to increased reliance on artificial intelligence (AI) coding tools, well over a third (35%) are either achieving daily or more frequent product deployments, with 36% deploying software multiple times per week. However, more than half (51%) also noted AI-generated code leads to deployment problems at least half the time.
Conducted by the market research firm Coleman Parkes on behalf of Harness, the survey also finds more than three quarters (78%) admit they have fragmented delivery toolchains, with 70% of respondents also conceding their pipelines are plagued by flaky tests and deployment failures.
More than three-quarters (77%) said teams often need to wait on others for routine delivery work before they can ship code and only 21% said they can add functioning build and deploy pipelines to an environment in under two hours.
Nearly three quarters (72%) also said they have hardly any standardized templates and “golden paths” for services and pipelines, and a similar percentage (72%) said they also believe their current ways of working are not sustainable over the long term.
Rahul Sood, general manager for Harness, said the survey results suggest that while developers are more productive in the age of AI software engineering, teams are also encountering existing bottlenecks in their DevOps pipelines more frequently. Many of those issues can be resolved if more tasks can be automatically assigned and completed both as code is being written and as it moves through a DevOps pipeline, he added.
A full 84% of respondents are already using AI tools daily for coding tasks, followed by quality assurance testing (68%), performance/cost optimization (63%) and refactoring (62%). Nevertheless, more than a third of a developers’ time (36%) is still spent on repetitive manual tasks, the survey finds.
At the same time, three quarters (75%) said pressure to ship quickly has contributed to burnout. A total of 71% of developers work evenings or weekends at least once per week because of release-related tasks or production issues.
It’s obviously still early days so far as adoption of AI coding tools is concerned, but as the pace at which code is being developed increases, any issue that already existed in a DevOps workflow is further exacerbated. In many cases, DevOps teams are encountering the same bottlenecks they also have, only a lot more often. For example, 86% of respondents agreed that security and compliance checks need to be more automated to meet delivery timelines.
Less clear is the impact AI will have on the size of software engineering teams. While some organizations have downsized the size of these teams because of anticipated productivity gains, many organizations are focused on enabling their existing teams to build and deploy software faster.
Regardless of how many software engineers might be employed, the one thing that is for certain is that software development in the age of AI will never be the same again. The only thing that remains to be seen is to what degree each individual developer instead of writing code becomes a software architect that oversees a small army of AI agents they employ.