Pushing the boundaries of AI in project management (while winning two APM Awards)

In 2023, when DHL found itself struggling to deliver its projects on time and to budget, it decided to rethink its approach to project management. It asked MIGSO-PCUBED to assemble and lead a team of data and project specialists to find ways to reduce the gap between its plans and what it delivered.

The resulting project, Intelligent Project Prediction (IPP), which combined the skills and experience of MIGSO-PCUBED, greyfly.ai and DHL, went on to scoop both the Technology Project of the Year Award and the Innovation in Project Management Award at the 2024 APM Project Management Awards.

Along the way, it demonstrated the potential value of artificial intelligence (AI) to the project profession.

Predicting project outcomes

“There’s so much opportunity for AI to enhance the project profession that it’s difficult to know where to start,” says James Martin-Young, Head of Digital and Data for MIGSO-PCUBED. “First, AI can enhance the use of data to predict outcomes. Second, generative AI and large language models will ultimately enable us to automate many lower-value tasks. This will enable project professionals to focus on higher-value thinking and may revolutionise how projects engage with their stakeholders and the organisations around them.”

The tool developed by the award-winning project – IPP – uses advanced analytics and machine learning to predict project outcomes, improving risk management and decision-making.

The project meant building a single integrated team, where representatives from DHL’s PMO and project teams worked alongside a delivery manager and experienced data, technical and change experts from MIGSO-PCUBED. The first task was for the MIGSO-PCUBED team to share their knowledge and experience of AI’s project portfolio management capabilities and to build a common framework that everyone could digest and discuss.

Two challenges: data and people

There were challenges. The first lay in the quality of the available data, which was initially too low to drive the necessary AI machine learning or algorithm training. The second was more traditional, and arguably more difficult to overcome: people at DHL were sceptical of AI’s ability to do the job, yet buy-in would be critical.

“Owners of an existing process, who may have had 30 years of delivery experience, were being asked to change,” says Martin-Young, “not because of a specific issue, but because an AI platform was saying it would be beneficial. We needed to convince them that their project may be at risk and that they simply weren’t able to see the patterns that show that.”

Martin-Young is quick to point out that, despite initial hesitancy, DHL “were ultimately one of the key reasons the project succeeded. We had a brilliant sponsor who was able to help guide other senior stakeholders through the ambiguity and help us shape what was required to satisfy their needs.”

It was also helped by deploying a “brilliant” AI change management framework, developed by the team to help accelerate the adoption.

Overcoming any lingering resistance required firm (but gentle) communication. The delivery team held regular meetings with the core representatives of the DHL project management community to ensure they understood the approach and its rationale.

How the team delivered the project

The team followed a core, six-step life cycle, focusing on the design, build and training of the AI model and its interfaces, followed by activities to understand the insights, identify focused and appropriate interventions, and finally deploy them.

Those initial interventions were inspired by data quality assessments and focused on supplying the AI model with good-quality data. And when the AI model was fully trained and providing quality project predictions, they focused on aligning delivery to budgets and reducing cost and time overruns.

Impressive benefits

The project lasted for 18 months and completed in January 2024. It doesn’t take computer hyper-intelligence to process the benefits. The level of data assessed as poor quality dropped from 78% to just 9%. Forecasting of a project’s over-spend reached 95% accuracy. This information helped the team reduce the proportion of projects exceeding budgetary targets from 33% to a mere 6%. Meanwhile, the average budgetary overspend dropped from 74% to 33%.

There were other knock-on improvements. Reports got better, as did the clarity of communication. Decision-making became more confident and focused. And the culture within DHL became more open to further AI initiatives.

Pushing the boundaries of AI

“We really have made a difference and I’m proud to have played a small part,” says Martin-Young. “I believe that the next phase of the IPP project at DHL will continue to push the boundaries of what AI can do to improve project outcomes. We’re on to a step change in the way data is used in the project profession.”

As for being named APM’s Technology Project of the Year, Martin-Young says it’s “all beyond words”.

“The recognition is great to have, especially among such stiff competition,” he says. “I’m just pleased that the hard work put in by everyone around the table has been recognised as a success.”