The project manager's dilemma was straightforward: The AI scheduling system flagged a three-week delay risk at a major university research facility based on historical data patterns. The experienced project team, drawing on institutional knowledge and stakeholder relationships, believed they could maintain the timeline. The algorithm disagreed strongly. Who should have final authority?
For most infrastructure organizations, this scenario triggers improvisation—urgent meetings, unclear accountability, inconsistent decisions that undermine both human expertise and algorithmic value. For Chirag Soni, it validated years of research into what he calls "the governance gap": the space between what AI can recommend and what organizations should authorize it to do.
Chirag Soni is an architect and project management expert specializing in AI governance frameworks for complex public infrastructure. After founding META Architecture and managing diverse commercial projects, he has supported delivery of critical infrastructure across National Air Guard facilities, hospital systems, and university institutions.
He holds PMI PMP certification, GBI Professional credentials, is a Licensed Architect in India, and an AIA Associate Member. His peer-reviewed publications appear in the International Journal of Technology, Management and Humanities and IEEE Xplore. He serves on editorial boards for international journals and has engaged with approximately 60 professionals through conference leadership and mentorship roles.
Soni's expertise in this space didn't emerge from academic theory. After founding META Architecture, where he developed comprehensive architectural and project management capabilities across diverse commercial projects, he moved into large-scale institutional infrastructure. His work on Clemson University's $130 million Advanced Materials Innovation Complex—an elite research university representing the highest tier of academic research activity—exposed the problem that would define his career.
91.5%
of construction megaprojects exceed budgets or miss schedules—a failure rate Soni attributes not to inadequate technology but absent governance frameworks
The Clemson project brought together three organizational cultures with fundamentally different operating logic: academic researchers working to scientific discovery timelines, corporate contractors operating by construction schedules, government stakeholders requiring strict compliance documentation. Digital twin technology could theoretically streamline coordination. AI predictive analytics could flag conflicts months ahead. But when the algorithm's recommendation conflicted with stakeholder expertise or institutional knowledge, nobody had frameworks for resolution.
"The governance gap became impossible to ignore when I realized project managers had no frameworks for deciding whether to override an AI recommendation. That moment shifted my career toward thinking systematically about this problem."
Rather than treating this as a project-specific challenge, Soni began researching structural questions: Why do projects repeatedly fail despite increasingly sophisticated tools? What governance frameworks ensure AI enhances rather than replaces human judgment? How do organizations deploy technology responsibly in infrastructure affecting national defense, public health, and educational missions—contexts where algorithmic mistakes carry public consequences?
From Practice to Research: Addressing Critical Gaps
Those questions drove him to publish peer-reviewed research over several years. His work in the International Journal of Technology, Management and Humanities addressed how project managers shift from reactive problem-solving to proactive risk management using AI predictive analytics. Research published in IEEE Xplore explored how digital twins create adaptive project environments enabling anticipatory management.
Additional research examined advanced building materials reducing installation complexity and lifecycle costs—directly connected to his current work directing mission-critical building envelope systems. His breakthrough contribution: "Human-Centered Governance Framework for AI-Integrated Digital Twin Systems in Construction Project Delivery."
Where most AI research focuses on algorithmic capability, Soni's framework addresses organizational authority: what humans should allow algorithms to do, and how organizations make that determination. The framework proposes explicit decision architectures answering critical questions before deployment: Which choices will AI influence? Which remain entirely human? Which require collaborative judgment?
"Most organizations deploy AI first, encounter problems, then scramble to create governance frameworks," Soni explains. "We need to reverse that sequence—establish governance first, then deploy technology within those frameworks."
Testing Frameworks Where Failure Isn't Abstract
Soni's current role at Roofing Solutions puts these frameworks to daily testing. Directing mission-critical building envelope systems across National Air Guard bases, hospital facilities, and university campuses means operating in contexts where system failure triggers security breaches, patient safety risks, or academic disruption. Each project requires comprehensive scheduling, budget management, procurement coordination, subcontractor oversight, regulatory compliance with federal and institutional standards, real-time monitoring, and documentation satisfying both government accountability and institutional requirements.
It's precisely the environment where governance matters most. National Air Guard facilities demand rigorous compliance verification where deviations trigger federal review. Hospital systems require rapid response capabilities where equipment problems affect patient care. Universities need flexibility for academic continuity while maintaining quality standards protecting institutional reputation.
His pending patent—"Advanced AI-Based Predictive Maintenance and Quality Monitoring System Using IoT Sensors, Computer Vision, and Machine Learning for Building Construction and Project Management"—emerged from managing these installations. Unlike monitoring systems that simply flag issues, Soni's framework integrates sensors, computer vision, and machine learning within governance structures respecting human expertise. The system identifies anomalies that might escape observation but positions final decisions with experienced professionals understanding specific organizational contexts—military security requirements, healthcare safety protocols, or institutional academic needs.
The Overlooked Asset: Organizational Memory
A critical focus of Soni's work addresses what he calls "the data blindness problem." Most infrastructure organizations possess decades of project history—schedules, cost overruns, risk events, stakeholder decisions, vendor performance, weather impacts, regulatory changes. This information exists in archives and institutional memory but organizations lack frameworks to leverage it strategically.
"Organizations are sitting on tremendous historical knowledge but have no systematic way to extract patterns, anticipate risks, or inform decisions based on what they've already experienced," Soni argues. "An institution that's delivered fifty projects over thirty years has invaluable data about what works in their specific context—their approval processes, stakeholder dynamics, contractor ecosystem, regulatory environment. That's competitive advantage if you can access it through big data analytics."
The practitioner-researcher duality: While directing infrastructure projects, Soni serves on editorial boards for the International Journal of Multidisciplinary and Scientific Emerging Research and International Journal of Innovative Research in Science, Engineering and Technology, evaluating manuscripts to ensure published research serves practicing professionals.
Building Professional Capability
Beyond his own work, Soni has committed to developing capability across the field. As Session Chair for the International Conference on Computational Technologies for Research in Data Analytics, he facilitated focused technical discussions among researchers and practitioners, connecting academic research with industry applications. As peer reviewer for the IEEE International Conference on Recent Trends in Computing and Smart Mobility, he evaluated ten manuscripts, assessing not just research rigor but practical relevance for project managers facing real implementation challenges.
Most directly, serving as judge for the OneEarth International Hackathon put him in conversation with eighteen emerging professionals addressing environmental and construction challenges using AI. "These professionals needed more than technical feedback," Soni reflects. "They needed context from someone who's worked across different delivery methods—understanding how innovations actually get implemented in organizations with competing priorities, regulatory constraints, stakeholder politics, and budget limitations."
Three Principles for Responsible Integration
From this accumulated experience—architectural practice, large-scale institutional infrastructure, research publication, professional development—Soni has distilled three pillars for responsible AI integration.
First, human-centered design. The most significant decisions involve contextual judgment no algorithm replicates. "The goal isn't automation," Soni insists. "The goal is augmentation. If we're replacing judgment, we're setting up for failure when AI encounters situations outside its training data—which happens constantly where every project has unique stakeholder dynamics, site conditions, and community impacts."
Second, transparent decision architecture. Organizations establish explicit frameworks before deployment: Which decisions will AI influence? Which remain human? Which require collaborative judgment? "Without answers established in advance, organizations scramble when conflicts arise, making reactive choices under pressure rather than following principled frameworks developed thoughtfully."
Third, accountability and learning loops. When AI recommendations diverge from human decisions, that divergence becomes data revealing where systems work, where they're misapplied, how they need evolution. "Deployment isn't completion," Soni explains. "It's the beginning of learning how your organization should use this technology given your unique context, constraints, and mission."
"The organizations leading infrastructure's next decade will be those asking: How do we design technology to augment human capability while maintaining human agency and capturing AI's analytical advantages?"
The Path Forward
Soni's vision is specific: Within five years, major infrastructure organizations will adopt governance frameworks before deploying AI, using decades of historical project data to inform decisions rather than reactive problem-solving. He's working toward this through direct project delivery—proving frameworks in mission-critical contexts where failure isn't abstract. Through research publication ensuring frameworks reach beyond individual projects to shape professional standards. Through professional development ensuring the next generation understands governance as enabler of responsible innovation.
His book, "Project Intelligence: AI-Enabled Decision and Execution in the Built Environment," synthesizes this knowledge into practical frameworks project managers can implement regardless of project type or organizational structure. The book offers adaptable templates organizations customize to specific contexts while maintaining core principles about human-centered decision-making and operational transparency.
"Public infrastructure delivery stands at an inflection point," Soni observes. "AI, digital twins, predictive analytics—they're being deployed today across project sites serving defense, healthcare, education. The question isn't whether technology transforms project management. It already is. The question is whether the transformation will be thoughtful or chaotic."
For organizations considering AI adoption, his message is direct: establish decision-making frameworks your teams understand and trust, then deploy technology in service of those frameworks. "The organizations leading infrastructure evolution aren't deploying the most sophisticated AI," Soni contends. "They're being most thoughtful about human-technology collaboration. That thoughtfulness is competitive advantage that's difficult to replicate because it requires expertise across design, execution, research, and governance."
His work demonstrates what's possible when governance frameworks precede technology deployment—where AI recommendations inform decisions without overriding human judgment, where organizational knowledge accumulated over decades gets systematically leveraged, where technology serves mission-critical infrastructure protecting public safety, national defense, healthcare delivery, and educational missions. The future is being built one thoughtful framework at a time—by professionals ensuring technology serves the public good rather than just demonstrating capability.
Connect with Chirag Soni:
