Published on the 01/05/2025 | Written by Heather Wright

Value-clearing opportunities across the value chain…

From virtual dispatcher agents and subject matter experts to genAI enhanced messaging platforms, generative AI is reshaping supply chains according to McKinsey, but the consulting firm warns it won’t all be plain sailing and companies will need to double down on both technology and talent to reap the benefits.

The company says genAI has the potential to create a ‘step change’ in supply chain efficiency, but with convincing supply chain leaders could be a challenge, with many already burned by digital investments which have failed to deliver real value.

“It’s important to recognise its limitations and ensure that it complements human expertise and traditional software solutions.”

Boston Consulting Group is also bolshy about the prospect of GenAI in supply chains.

It says companies have been slow to realise the vision of AI-driven supply chains, largely due to an overemphasis on AI’s analytical powers, and little focus on applications which include adaptive learning.

“AI solutions, laden with complexity, often overwhelm supply chain personnel, leading to low adoption and diminished returns on investment,” BCG says.

GenAI, with its user-friendly interfaces, agent-base automation and cross-functional orchestration, could change that.

Alberto Oca, McKinsey partner and coleader of digital warehousing in North America, says GenAI is already revolutionising the logistics industry, claiming some big figures when it comes to boosting performance and operations – including US$18 billion in supply chain operations.

“GenAI offers value-clearing opportunities across the entire logistics value chain,” he says.

That ranges from core operations such as planning, optimisation, warehousing, transportation and asset management, to support functions such as procurement, customer experience, back office, legal and HR.

“GenAI enhances various aspects of the value chain and leverage advanced AI models to create new, content-optimised processes and improve efficiency across logistics operations.”

While reducing administrative burden through assistance with documentation is a common use of genAI – and in the case of supply chain can reduce  lead time for producing documentation by up to 60 percent according to McKinsey, through auto-generating and consolidating shipping documents and identifying potential mistakes and ‘digesting’ corrections – agents are bringing with them new opportunities.

Oca cites the example of virtual dispatcher agents which work alongside dispatchers, assisting drivers with troubleshooting and roadside assistance.

He says one last-mile operator with a 10,000 strong vehicle fleet saved $30-$35 million its $2 million investment in virtual dispatcher agents.

Virtual subject matter experts which can synthesise findings from multiple systems to provide answers for managers are also available, while three-way messaging platforms leveraging genAI and standard SMS text messaging are connecting drivers, dispatchers and customers in a single text conversation to improve efficiency and ‘instantly’ resolve delivery-related issues.”

BCG says there are four levels of genAI capabilities for supply chain operations – ranging from the basic deployment of task-specific point solutions, to those that can reshape the industry through process enhancements, deep process transformation and cross-functional process automation.

Process enhancements sees genAI acting in combination with existing planning and execution systems to improve the effectiveness of current processes, such as monitoring supply chains for disruptions, generating alerts and simulating responses. These products are already available, BCG notes, with more in development.

Deep process transformation sees genAI agents continuously verifying and updating master data sets and ‘driving rethinking of entire workflows’ transforming core processes.

BCG says to fully unlock the potential of genAI agents here, ‘significant’ process re-engineering will likely be required, in many cases necessitating custom built solutions or at least tailored add-ons to ensure genAI is seamlessly integrated.

The most advanced level of in BCG’s model is cross-functional process automation, where groups of self-organising genAI agents will orchestrate supply chain operations across different functions creating an automated, intelligent and collaborative system.

But McKinsey’s Oca says genAI in the supply chain doesn’t have to be complicated or a huge investment in order to reap returns.

“The focus should be on creating a strong foundation that companies can build on. The most important thing is to define that foundation and how you can expand on it.”

He cautions companies to also think about the workforce of the future, saying the adoption of genAI in logistics can change the workforce dynamic, shifting away from manual workload and more toward value-adding tasks – requiring a shifting of gears for employees.

“There’s going to be a lot of scepticism, so you’ll need to quickly prove the technical feasibility and estimate the business value. Then you can start developing minimum viable products and scale up across the business.”

Asaf Somekh, cofounder of McKinsey’s Iguazio AI platform, urges against ‘one-off attempts or MVPs’. Harking back to Oca’s comments about building the right foundations, Somekh warns that many clients have build genAI applications they were satisfied with only to realise running them at scale is too high to justify the value they offer.

“This is where many projects get stuck. The right architecture needs to be in place, but there’s a big gap in the industry when it comes to how to use those GPUs [on which genAI runs].

“It’s one thing to get the application to work as intended, but quite another to do so cost-effectively.”

Somekh is also clear: Not everything needs to be genAI-based.

“GenAI is expensive in terms of compute resources so we always recommend people set up a flexible environment that combines traditional AI, like machine learning and deep learning, with genAI,” he says.

“Machine learning is good enough for some things.”

Adds Oca: “Similar to automation, it’s important to recognise its limitations and ensure that it complements human expertise and traditional software solutions.”