Project management without the hype: How PlanAutomate uses AI for real results

This article is sponsored by PlanAutomate.

Managing complex projects in a fast-moving organization is never simple, especially when new technology, such as AI, is expected to solve every problem. 

During a recent MSDW webcast, Matt Mong, chief category officer at PlanAutomate, set out to address what really works today — and what doesn't — when it comes to integrating AI into project management duties and workflows. Real-world examples are critical for organizations trying to evaluate their options today, he explained, because they provide perspective on what processes and operational areas are good candidates for AI-based solutions today.

What AI can and can’t do in project management

Generative AI can handle simple creative tasks, but when it comes to areas that demand total consistency, such as moving money, assigning resources, or keeping projects on track, the reality is that humans and purpose-built software are still necessary.​

But when faced with project data at scale, AI tools can prove valuable.

"AI is brilliant at one thing, and that is analyzing massive, complex data sets to find patterns that typically no human can find, at least not easily," Mong said. But to truly benefit from these insights, project teams have to break down silos and bring together information about finances, resources, supply chain, and risk management. 

PlanAutomate's model is built on this kind of unified project data. Mong described how bringing a project's moving parts into one source, such as Microsoft Dynamics 365 Finance, allows AI to monitor and interpret changes in real time, from budget shifts to resource conflicts and supplier delays.

"You need to natively unify all the critical project data and processes, so the financials, the operations, your resources, supply chain, your risks," he explained. "All those things need to be in a central location, one single source of truth to create this unified data set. And this is really that secret ingredient."

How PlanAutomate’s AI surfaces risks and recommends next steps

During the webcast, Mong demonstrated how users can interact with Plan Vector, the AI-powered assistant inside PlanAutomate. Instead of relying only on dashboards, stakeholders can ask Plan Vector for a project status report or drill down on specific problems, such as supply chain delays or resource overloads. 

Plan Vector identifies root causes and suggests possible actions through a chat-based interface. 

Mong detailed how their AI is trained on synthetic project historical data, learning causality and relationships. This makes the solution feel like having an experienced project manager or project director who understands why a project might go off track, not just that it's happening. 

"We show the AI hundreds of years of project lifecycles. And what it learned during that training is essentially causality," he said. "So it basically understands how projects move, how to interpret the data that you're showing it, how to tease out those subtle issues in the data before they become big problems downstream that really threaten the success and threaten to take your projects to failure."

Mong explained that PlanAutomate’s platform brings together project scheduling, resource management, and supply chain information within a single system. During the demonstration, he showed how users can see resource status, for example identifying when a team member is overloaded or when supply chain delays occur. 

The AI highlights these risks, showing which tasks or resources are affected, and suggests certain actions, such as rebalancing your resources or addressing the schedule or supply chain issues. "It's going to give you some suggestions as far as how to fix this project," Mong said.

Why full automation is still a long way off

But Mong also warned against expecting too much too soon. Full automation in project management isn't realistic yet, especially for high-stakes projects or high-stakes business. Trust needs to be earned gradually through transparency and solid results. 

"You can't really have a reliable AI automation without first having this experience, without first having it be able to understand what is going on in your projects, just like a human would," Mong explained. 

For now, tools like PlanAutomate aim to boost insight and analysis, providing what Mong called a veteran analyst that helps project teams see problems early and take smart action.​​

The webcast wrapped up with practical advice. "The most powerful use case for using AI is to automate that insight first. Automate the analysis," Mong said. He urged project-focused organizations to prioritize building a solid, unified data platform before chasing more advanced AI-driven automation. 

With stakeholders able to ask the system questions, drill down into data, and get recommendations, Mong predicted that eventually, as AI proves itself, organizations will move from human-in-the-loop decision-making to trusting the system to make certain changes independently.