How GenAI Is Reshaping Supply Chain And Operations (2025)

Alain Gagnon, Senior Director at PwC and Microsoft Lead for Eastern Canada, contributed to this story.

Generative AI (GenAI) has been a hot topic for years across industries as organizations and business leaders strive to implement these technologies to maximize its benefits.

Thus far, the implementation of GenAI for most organizations has centered on administrative activities and use cases, which are useful for shortening tasks like taking meeting minutes, translating documents, or creating agendas.

Leaders should look beyond the desktop use cases and consider applying GenAI to areas like Cost of Goods Sold (COGS) and Operating Expenses (OpEx). These functions could reveal additional hidden value through the use of data-driven optimization. So far, relatively few companies have effectively enabled the technology and processes required for utilizing data to yield the most benefits for their organizations by deploying GenAI solutions.

Organizations can gain significant competitive advantage by being early adopters to this form of GenAI implementation in their supply chain and operations. Some clients have seen up to 10 to 15% savings on their project costs and/or time after integrating Generative AI into their supply chain operations.

These savings can then be redirected from non-value-added tasks to more strategic activities, creating value and enabling companies to scale. Given the current economic uncertainty amid looming tariffs and pricing pressures, a reduction in costs or time can free up an organization’s resources for innovation, strategic initiatives, and overall efficiency improvements.

In working with clients, we have found there are four main roadblocks most organizations face when trying to achieve tangible results with GenAI in their supply chain and operations:

• Data Issues: Outdated and unreliable data
• Process problems: Immature and undocumented processes
• Legacy systems: Limitations of legacy systems and ERPs
• Stakeholder alignment: Misaligned stakeholders or lack of a clear roadmap and program

To overcome these barriers, organizations should keep in mind a few key tips.

Fail Fast

Many organizations sit on decades of unstructured – and sometimes poor-quality – data. The prospect of cleaning it all in a linear and structured approach, which could take years, can be overwhelming, leading many to delay, or avoid GenAI initiatives altogether.

Instead, organizations should adopt a fail fast approach. This strategy enables companies to organize data in a way that creates a GenAI foundation, thus being able to go live in weeks rather than months or years.

As such, a subset of the data identified as most relevant is selected as the foundation data set. We then use our model to test against it, iterating quickly to identify inaccuracies. This process is crucial – if you do not succeed in establishing an accurate data product, models will not yield useful results. Therefore, speed, adaptation and precision are key.

Improve, Don’t Replace

Many organizations want to incorporate new technologies into their operations but get stuck thinking they need to swap out their legacy systems first. This is not true as organizations can work with these systems instead of against them.

For example, APIs can extract data from legacy systems and databases to feed data products to make the models more intelligent and flexible for clients.

Start Small, Be Targeted

Many organizations attempt to tackle a large portion of their operations at once. Instead, it’s advisable to go begin with specific, well-defined use-cases in operations, supply chain and manufacturing where data is readily available to be able to map the current state of the process and future state of technology.

A good rule of thumb is to identify use cases that have some available historical data, are well described and show clear potential to reduce costs and accelerate processes or remove manufacturing roadblocks.

We use iterative deployment approaches to help clients experience the actual tools to operate their business without having to invest in new systems or large modules like SAP and Oracle. Low code and no code applications are preferred to unlock the value of the GenAI models developed.

Consider the Whole Roadmap

Having a robust GenAI strategy and roadmap is essential to long-term success. Equally important is having a delivery team that can bridging strategy and implementation.

Before embarking on any project, it is essential to ensure your team possesses the right capabilities to support both. Organizations benefit most when they partner with firms that offer both strategic planning and hands-on GenAI execution experience and expertise. Having access to both skill sets is key for success.