Shortcut this week added an artificial intelligence (AI) agent that is capable of orchestrating the planning, tracking and coordination of software projects.
Company CEO Kurt Schrader said the Korey agent takes ideas described in natural language and converts them into structured, build-ready plans in seconds, including generating detailed specifications with clear acceptance criteria. It then breaks the project down into a set of actionable tasks, tracks dependencies and summarizes work status.
It is designed to pull context from projects, comments, GitHub activity and other connected tools to ensure plans are accurate, complete and actionable. As Korey continues to learn a team’s preferences and workflow style, it functions like a product manager that enables software engineering teams to better keep track of the actual progress they are making toward completing a project, said Schrader.
DevOps teams, as a result, can gain instant visibility into projects, including any blockers that have emerged. They can also modify assignments, set up sprints, move items across workflows and organize epics to manage projects. Specific tasks can then be delegated to engineers, designers, or even other AI agents.
The overall goal is to make it simpler to create, modify and organize projects by, for example, eliminating the need to manually create a story for each project, noted Schrader.
Once known as Clubhouse, Shortcut launched a project management application in 2014 that provides an alternative to Jira for managing software development projects. Existing customers include The Farmer’s Dog, Dataiku, Crossbeam and Octopus Deploy. Longer term, the company now plans to also extend the reach of Korey to both Jira and other tools and repositories that software engineering teams have historically relied on to manage projects, said Schrader.
It’s not clear how rapidly software engineering teams are moving beyond AI tools to write code to embrace AI agents that organize workflows and automate tasks, but much of the toil that today conspires to slow down the pace at which applications are built and deployed has little to do with the writing of code. As more mundane tasks are automated using AI agents, it should become easier for DevOps teams to realize the productivity benefits that AI technologies are capable of providing.
In the meantime, however, DevOps teams would be well advised to start identifying mundane tasks that might soon be better performed by an AI agent. It’s not likely that AI agents will replace the need for software engineers any time soon, but the amount of time spent on tedious tasks should be sharply reduced in a way that ultimately reduces the level of burnout most DevOps teams currently experience.
Of course, it may take a few months yet to fully realize that potential, but the issue now is not so much whether to rely on AI so much as it is determining to what level it can be trusted to autonomously complete a task. The one thing that is for certain is that the only ones who will know for sure if something has been done right in the first place are the humans, who will ultimately be held accountable for ensuring the success of any given project.
