It is hard to name all applications of AI in release management when all software tools used in a release start having AI components here and there, so the safest is to list some release stages where AI is having the biggest impact at the time of writing. According to my experience these are Release Management, Development, DevOps and Quality Control.
The current minimum level is using an AI release coach, an agentic component.
Uncontrolled usage: According to my experience most release managers use AI in an uncontrolled form: privately accessing a public AI chatbot for checking their task assigments and risk assessment lists. Basically the modern form of a checklist, like looking up the basics in a book.
Specialized AI Release Coach: Beyond generic AI chatbots there are specialized AI Release Coach services. An AI agent can be regarded being specialized if it is prepped up with release related knowledge. When certain release managment knowledge base is connected to the AI agent and that knowledge is constantly being updated by industry best practises of release management methodologies. more...
AI Release Coach with HITL: HITL in AI means Human-In-The-Loop. When a specialized AI Release Coach has certain approval points built in, where senior human release coach can give approval for decisions and even for specific recommendations. IT Releases are usually very expensive processes involving dozens of high-paid professionals from the whole spectrum of IT. Days of delay or failed GOLIVES can easily cost a company 6 or 7 digits. This is the reason why AI needs professional human oversight in certain cases. more...
There is no need to explain how deep AI has gotten into code development. We can safely say that today no software development can be operated efficiently without AI code tools.
Probably CI/CD and DevOps in general was the first to apply highly automated practices with deploy pipelines and related package managment. AI definitely made DevOps quicker and cheaper.
AI put a relatively high pressure on testing. The speed at which AI can create code is changing the pace software testing (design and execution) was structured. But by now it is not only the amount of testable code from AI that impacts Testing but test tools have also started to use AI enhancements. Generating test cases or test automation scripting can hardly operate without the help of AI. Though it is important to mention that at the moment AI cannot be left alone with testing. HITL is crutial to ensure that test results, coverage, and pass rates are reliable. I tend to agree that visual test automation is the proper answer to AI generated code testing. (see more about this topic here)