The pace at which multiple classes of artificial intelligence (AI) are converging to automate DevOps workflows is about to accelerate the rate at which higher-quality applications will be both built and deployed. And this will happen at levels of scale once thought to be unimaginable.
Most recent AI advances have been squarely focused on using large language models (LLMs) to improve the productivity of developers. As valuable as that use case for AI may be, however, the volume of code moving through a continuous integration/continuous delivery (CI/CD) platform is only going to increase exponentially as developers become more productive. Most organizations are not going to be able to hire additional software engineers to manage additional pipelines, so the need to apply AI to DevOps workflows is becoming more pressing. Each passing day, more code is generated using tools that have been augmented in a way that makes it simpler to build higher-quality applications.
AI makes it possible to analyze large amounts of data in real-time to surface insights and optimize workflows. For example, the data analyzed by algorithms can identify where there are bottlenecks in a DevOps workflow, along with recommendations for how to break that logjam. DevOps teams can then apply machine learning algorithms to optimally configure a platform, otherwise known as AIOps, in a way that eliminates many of the manual tasks that make managing DevOps workflows more tedious and time-consuming than anybody wants. AI is also serving as a digital ‘pair programmer’ for many DevOps professionals, helping them think through problems and brainstorm innovative solutions quickly.
Generative AI tools go a step further—using natural language to instruct an LLM that has been trained to generate code to create the scripts that automate DevOps workflows. Those scripts can then be applied using an orchestration engine to automate CD processes at levels of unprecedented scale.
At the same time, organizations will also be able to use technologies such as a vector database to compare code against examples of other code that an LLM has been trained to recognize, with the goal being to identify defects in newly written code that might adversely impact application security. It’s now only a matter of time before the overall quality of the code finding its way into production environments improves to the point where the number of incidents a DevOps team has to respond to should start to decline steadily.
Realizing those benefits and other advances will require multiple types of AI models to be built and maintained. Companies such as CloudBees are making investments to define machine learning operations (MLOps) practices to build them. There will also be untold numbers of LLMs built by the open source community and enterprise IT organizations that will need to be integrated into DevOps workflows. The ability to review those models in a way that promotes transparency and trust will, of course, be crucial. As such, the value of an open CI/CD platform like Jenkins that makes it possible to embrace multiple classes of AI models simultaneously has never been more apparent.
Naturally, there is as much fear and trepidation about AI as there is excitement. No one knows for certain to what extent DevOps workflows might be automated, but there will always be a need for DevOps engineers to validate them. After all, it’s one thing to be wrong. It’s another thing to be wrong at scale. Mistakes that occur because of a flawed AI model could turn out to be catastrophic.
However, there is no going back. As multiple classes of AI technologies converge, DevOps teams will find can more easily manage DevOps workflows at scale. In fact, one of the long-standing criticisms of DevOps, in general, is how difficult it can be to manage workflows at scale. AI advances will soon effectively render that debate moot.
In the meantime, DevOps teams today would be well-advised to start reviewing their current workflows with an eye toward determining which processes can be automated. AI will soon be pervasively applied to drive a new era of software development that promises to dramatically accelerate the rate at which better applications can be built, deployed and updated. In fact, DevOps itself, thanks to the rise of AI, will soon be democratized in a way that will enable more organizations to embrace it, mainly because the level of programming expertise required will be significantly less.
As these changes occur, the number of applications that organizations can cost-effectively employ to drive an even wider range of digital processes is only going to increase. Exactly what impact all that software is about to have on the world as we know it remains to be seen, but the one thing that is for certain is that DevOps engineers will be at the center of it.
