An interview with Christian Pallaria, Digital Engineering Lead, and Sydney Mudau, PMI Sydney Chapter President, Principal Project Manager.
Too often, PMs find themselves consumed by daily technical issues when, instead, they could leverage AI to alleviate these burdens and direct their attention to more critical aspects that might otherwise go unnoticed and discovered too late.
Introduction
Through this article, Christian and Sydney examine how AI processes structured and unstructured data, integrates with different project management methodologies, and evolves into autonomous agents that project managers can use to enhance efficiency and strategic decision-making.
Project management has evolved significantly over the years, moving from manual scheduling and documentation to digital platforms that streamline workflows. However, construction project managers still face persistent challenges: delays, cost overruns, risk management and information overload.
With the rise of Generative AI (GenAI), a new era of project management is emerging. AI is not just a tool for automation; it is becoming a decision-making assistant, helping project managers analyse risks, optimise schedules and streamline operations. However, AI adoption in PM remains inconsistent, with many organisations unsure of how to leverage AI agents effectively.
The evolution of project management
Traditionally, PM relied on paper-based documentation, manual scheduling, and in-person communication. As projects became more complex, software tools like Primavera P6, Microsoft Project, and BIM platforms were introduced to improve efficiency. Despite these advancements, construction projects still face significant challenges:
- Complexity — managing multiple teams, vendors, and stakeholders across large-scale infrastructure projects
- Delays — supply chain disruptions, weather conditions, and unexpected design changes
- Cost overruns — inefficient resource allocation and poor risk management
- Scheduling challenges — balancing unforeseen site conditions and scope changes
- Data overload — vast amounts of structured and unstructured data that are difficult to turn into meaningful insight
With these challenges in mind, AI is emerging as a game-changer, enabling PMs to automate workflows, predict risks, and make data-driven decisions with greater precision.
From chatbots to intelligent agents
On November 30, 2022, OpenAI released ChatGPT, ushering in a new wave of AI-powered tools. Since then, AI has evolved from simple chatbots to intelligent AI agents capable of real-time decision-making. OpenAI's File Search and Agent SDK significantly enhance how AI integrates into PM: File Search enables real-time semantic search across project documentation, transforming contracts, meeting minutes and emails into instantly searchable knowledge bases. Agent SDK provides a framework for building autonomous agents tailored to construction workflows, offering PMs real-time strategic decision-support — effectively an intelligent "Deputy PM."
Today, AI in PM is used primarily for task automation, risk assessment, and resource optimisation. Soon, we'll see AI agents acting as Deputy PMs that detect project risks before they arise, recommend budget adjustments and procurement strategies, and continuously assess project health in real time.
The AI Agent Era has arrived. The key challenge is how to integrate AI effectively into PM workflows. PMs don't need to fear this shift — as has happened with new technologies throughout past industrial revolutions — but should instead learn to communicate and coordinate their work with generative AI. The key sits in the interaction between humans and GenAI, rather than competition.
Structured vs. unstructured data
The main technical obstacle is the very large amount of unstructured data that every company holds. The effectiveness of AI in project management depends on data availability and structure. Structured data — schedules, budgets, material quantities — is ideal for predictive models like machine learning. Unstructured data — emails, meeting notes, contracts, handwritten reports — is difficult to analyse using traditional tools.
Most construction firms have large amounts of unstructured data but lack the tools to extract insight from it. This is where GenAI offers a breakthrough, transforming unstructured data into structured formats that AI can analyse — extracting key clauses from contracts, flagging critical action items in emails, and synthesising scattered project data into coherent, actionable reports.
Waterfall vs. Agile: tailoring AI to different PM approaches
AI can enhance both traditional (Waterfall) and Agile project management, but its role varies with the approach. In Waterfall PM, AI supports predictive scheduling, risk analysis, and automated reporting — for example, processing a five-year schedule to flag tasks at high risk of delay, or cross-referencing ACWP, BCWP and BCWS to highlight tasks that are over budget or behind schedule and propose alternative solutions. In Agile PM, AI supports resource optimisation, sprint planning, and continuous feedback loops — tracking sprint performance and recommending real-time adjustments to task assignments. In both cases, the model becomes a genuinely custom assistant when trained on the specific project, client, and resources, rather than a generic tool.
Conclusions
To stay competitive, project managers must start integrating AI today. Custom GPT models already offer real-time project analysis, risk prediction, and workflow automation. Looking ahead, the next generation of AI-driven PM applications will minimise human intervention by automating up to 80% of routine tasks, provide predictive insights to prevent cost overruns and schedule delays, and adjust decision-making strategies dynamically based on project conditions.
AI is no longer a futuristic concept — it is a critical tool for project managers today. Those who embrace AI-driven project management will lead the industry, while those who resist it risk falling behind. The future of project management is AI-assisted, data-driven, and increasingly autonomous. The key is to start integrating AI today, before falling behind becomes the real risk.