How can analyzing digital project workflows help managers predict risks and improve project success?

Host: Welcome to A.I.S. Insights — powered by Living Knowledge. I'm Anna Ivy Summers, and today we are looking into a fascinating new study titled "Process Mining for IS Project Success Factors Management: A proposal." At its core, this study proposes a decision support framework that brings process mining into the realm of information systems project management. By analyzing the digital event logs left behind in our everyday project tools and code repositories, the authors outline a way to track workflows, catch deviations, and connect operational performance directly to critical project success factors. Joining me today to unpack all of this is Alex Ian Sutherland. Alex, great to have you here.

Expert: It is wonderful to be here, Anna. This study tackles a challenge that virtually every technology leader has wrestled with: how to truly understand whether a software or IT project is drifting toward failure before it is too late to fix it.

Host: Exactly. And that brings us to the big problem highlighted in the study. For decades, researchers and consultants have produced long lists of critical success factors—things like top management support, clear requirements, user involvement, and effective planning. Yet, despite knowing these factors, IT projects still run over budget, slip past deadlines, or fail completely. Why hasn't all that theory translated into daily practice?

Expert: That disconnect is precisely what the authors zero in on. Traditionally, success factors are treated as static concepts. They are listed in academic journals or reviewed informally during post-mortem meetings long after a project has wrapped up. In the real world, project managers rarely have concrete, data-driven ways to monitor these success factors continuously throughout the project lifecycle. Instead, management tends to be intuitive or reactive. When a task is skipped or a stage takes twice as long as intended, managers often have no immediate way of assessing how that specific deviation threatens their overall success factors.

Host: So how does this study propose changing that dynamic?

Expert: The authors—Joana Pedrosa, João Varajão, Luís Magalhães, and Ricardo Martinho—ground their proposal in Design Science Research. Their approach uses process mining to turn project execution into an objective, data-rich diagnostic and forecasting tool. Every time a software team updates a ticket in Jira, moves a card in Trello, or pushes code to a Git repository, they leave digital footprints known as event logs. Process mining ingests these logs to reconstruct the actual workflow as it happened, rather than relying on how someone hoped it would happen.

Host: That makes a lot of sense. So instead of guessing how the team is working, the system sees reality. Walk us through how their proposed architecture operates.

Expert: The authors design a two-phase architecture: diagnosis and prognosis. In the diagnosis phase, raw event logs from disparate tools are gathered into a centralized project log repository, formatted under standards like the eXtensible Event Stream, or XES. From there, automated process discovery and conformance checking analyze what occurred. It measures core performance metrics—such as lead time, service time, waiting time, and synchronization delays—while comparing real activities against the prescribed project management model.

Host: And that lets them identify when something goes off script, like skipping a formal review or an essential testing step.

Expert: Exactly. And here is the real breakthrough of the study: they introduce a business logic engine that links those operational deviations directly to the project's success factors. Using decision mining techniques, such as decision tree learning, the system can determine how skipping a specific task or lingering too long in a specific stage correlates with negative outcomes on known success factors.

Host: Which brings us to that second phase you mentioned—prognosis.

Expert: Precisely. Once the diagnostic engine identifies historical patterns and rules, the prognosis phase steps in. It evaluates the current trajectory of an ongoing project against historical data from similar initiatives. If it detects a deviation that historically degrades a key success factor, it triggers an alarmistic notification and visualizes the risk on a management dashboard. More than just ringing an alarm, it provides prescriptive guidance, suggesting adjustments to current and future tasks to bring the project back in line.

Host: Let's talk about why this matters so much for business leaders. For an executive or a PMO director, what is the practical impact of moving from traditional project management to this kind of process-mined framework?

Expert: The primary business value is transforming project governance from an autopsy into active, real-time life support. Typically, executives discover cost overruns or scope creep weeks or months after the damaging behavior took place. By continuously mining operational data, this framework provides early warning indicators. It removes the subjectivity and optimistic bias that often distort weekly status reports. If team members are circumventing critical quality gates, leadership sees the operational drift immediately and understands its exact risk to delivery.

Host: It essentially takes the guesswork out of project health and gives teams a chance to course-correct while there's still time and budget left.

Expert: That is the ultimate goal. It bridges the gap between daily developer activities and executive-level strategic success factors, creating a truly transparent, self-learning organization.

Host: It is a genuinely promising direction for managing complex digital initiatives. To recap, this study introduces an innovative, two-phase framework using process mining to discover workflows, link task deviations to success factors via decision mining, and deliver predictive alerts to project leaders. Alex, thank you so much for breaking down this research for us.

Expert: My pleasure, Anna. Thank you.

Host: And thank you to all our listeners for tuning in to A.I.S. Insights — powered by Living Knowledge. Until next time, keep seeking deeper insights.