In a bustling restaurant kitchen, efficiency requires more than just machines that wash dishes or chop vegetables. It requires a conductor to ensure the appetizer, main course, and dessert are prepared in the right sequence, that the right chef gets the right order, and that the correct dish reaches the right table on time.
The same dynamic applies to building software. When developers across the globe collaborate on open-source projects, writing the code is only half the battle. Coordinating the work is often the real bottleneck.
That coordination challenge sits at the center of new research co-authored by Alan (Ao) Huang, a third-year doctoral student at the University of Miami Patti and Allan Herbert Business School, alongside his advisors Ni Huang, professor and the Dennis and Smith Family Endowed Chair of Business Technology, and Yili Hong. Published in Information Systems Research, the study examines how workflow automation speeds up innovation in open-source software (OSS) development and finds that the type of automation used matters just as much as the automation itself.
The paper is Alan Huang's first publication in Information Systems Research, one of the field's most prestigious journals.
The researchers analyzed more than 4,500 repositories on GitHub, where developers collaborate on software projects, and 280,000 development issues. They found that workflow automation reduces issue resolution time by 10.1 percent, saving an average of 4.3 days per issue. At scale, the impact is significant. Assuming approximately 1.5 million active projects on GitHub and a 30 percent adoption rate of these tools, the researchers estimate that workflow automation translates to roughly $254.2 million in monthly labor-capacity savings across the platform.
"Those time savings mean less waiting, fewer manual handoffs, and more time for value-creation work," said Ni Huang. "Developers can spend less time managing repetitive process steps and more time solving unique technical problems."
Mechanization vs. Orchestration
The study introduces a critical distinction between two types of workflow automation: mechanization and orchestration.
Mechanization is the equivalent of the kitchen's dishwasher. It takes a repetitive, well-defined task and executes it reliably and quickly. In software, this means automatically testing code, checking formatting, or compiling builds. The research shows mechanization is highly effective for routine maintenance, saving an average of three days per issue.
Orchestration manages how people, tasks, and information move together. It automates the coordination between people and tasks. When a new feature involves multiple modules and requires contributors with different expertise, orchestration automatically labels the issue, notifies the right maintainers, triggers security checks, and assigns follow-up tasks, keeping everyone updated throughout the process.
While mechanization handles the routine, orchestration is what drives substantive, creative innovation. The study found that adding orchestration to the workflow saves an average of 9.1 days for complex new development.
"Mechanization is useful for routine maintenance because it is good at execution," Huang said. "But for creative, collaborative innovation, orchestration is important because it helps people, tasks, and information move together. It reduces the friction of collaboration, which becomes especially important in large-scale, distributed innovation environments."
Speed without sacrificing quality
A common concern in software development is that moving faster leads to rushed, lower-quality work. The data showed the exact opposite.
Projects using workflow automation not only resolved issues faster, but they also saw higher contributor engagement and better innovation outcomes, including more closed issues, more community stars, more forks, and more releases.
"Faster coordination in OSS can help sustain contributor engagement, maintain project development momentum, and create a more vibrant development community," Huang said.
Looking ahead
The findings offer a clear lesson for any organization trying to innovate faster: automating repetitive tasks is a good start, but automating coordination is where the real leaps in speed occur.
The research team is already looking toward the next frontier: AI coding agents that can actively participate in repository-level development work.
"We are interested in a larger economic and societal question," Huang said. "If AI can reduce the technical and coordination barriers to software creation, are we entering a world where software development becomes significantly more democratized? It is raising some very interesting questions about the future of software innovation, participation, and governance."
