Smoothing out the supply chain

Automation within the supply chain is growing, with each aspect, including sourcing, manufacturing, order and inventory management, sales and delivery of product, reaping the benefits of continued innovation.

Some of these benefits stem from robotic process automation software, or RPA. For many manufacturers, RPA bots now handle repetitive tasks such as filling in forms, funnelling sales inquiries, moving files and more – and they do it all around the clock.

Currently, roughly 50 to 60 per cent of large manufacturers are now using RPA, according to Daniel Cascone, vice president and senior partner at IBM Consulting Canada, and Marc Haesen, partner at the Global Industrial Manufacturing Center of Excellence, IBM Consulting. And while the adoption rate among Small & Medium Enterprises (SMEs) is significantly lower, with IBM reporting only about 10 to 20 per cent of SMEs utilizing the technology, that is expected to change.

This is because, like other simple software systems out there, RPA bots are now being employed in conjunction with AI. Cascone and Haesen report that AI technology in supply chain management is now more accessible than RPA and, because of AI’s ease of integration and broad applicability, “SMEs in particular may embrace AI faster than RPA”, Cascone reported.

In short, as AI’s power grows, the ROI of automating more supply chain management is rapidly coming within SME reach, allowing these firms to leapfrog over the RPA stage to attain a wider array of supply chain automation benefits.

The AI difference
Through aspects like machine learning, natural language processing and computer vision, AI greatly surpasses simple rule-based RPA functionality. RPA carries out straightforward defined tasks, while AI systems, called AI agents in the supply chain context, examine the data provided to them for patterns, irregularities and more. As they learn over time, AI can change the code of RPA bots so they’re not only better able to handle tasks, but they can take on more tasks and tasks with greater complexity. This represents a true paradigm shift, allowing companies to automate not just tasks but entire processes.

However, Amber Salley, VP of industry solutions at GAINSystems, cautioned that while these entities decide and act, they don’t replace human expertise, only augment it. “Human surveillance is still essential to validate assumptions, avoid scope creep and ensure output fidelity,” she explained. “While the potential is transformational, this new model requires trust, transparency and grounded implementation.”

Biggest impact
As to where these systems will make the biggest improvements in supply chain management, Cascone and Haesen first flag customer service. “AI agents are redefining customer service and operations by delivering a new, seamless digital experience for both customers and employees,” Haesen explained. According to a study by the IBM Institute for Business Value, organizations employing AI agents in customer service have seen a 60 per cent improvement in first-contact resolution.

AI agents will also have a large impact on supply chain management in general proactive monitoring and issue detection across a range of areas. “In the long term, AI will play a pivotal role in monitoring KPIs and other key metrics in real time,” said Haesen. “It will proactively alert teams to potential issues well before they become apparent, enabling faster decision-making and reducing disruptions.” He added that, “as AI adoption matures, it will help address complex data quality issues that are difficult to handle manually. By improving data accuracy, companies can enhance operational efficiency and achieve deeper optimization across their supply chain.”

Barriers to adoption
As is the case with implementing any new automated system, company leaders need to ensure their workforces transition effectively as AI and RPA take over some of their supply chain management work. In practice, Salley said this means managers will need to reframe roles, moving employees from transactional tasks to outcome ownership. “Teach teams to interpret AI outputs and model what-if scenarios,” Salley said. “Also, create review loops. That is, build structured processes for validating AI-generated actions and refining models through human judgment.”

In Cascone’s view, an effective initial approach is to provide a select employee with AI tools. “Encourage someone to start using it and demonstrate where these tools are beneficial and where human supervision is still necessary,” he said.

Cascone and Haesen recommend asking two key questions. “First, what repetitive tasks require significant staffing and demand more attention and effort than desired?” said Cascone. “By augmenting these tasks with RPA and AI, employees can focus on more strategic tasks and productivity will improve.”

Second, look at tasks that your organization should be tackling to streamline management of the supply chain, but has not prioritized due to the high level of resources required. “By addressing these two questions organizations can unlock opportunities to streamline operations, empower employees and focus on high-impact initiatives,” said Haesen.

Looking ahead
Although many AI supply chain applications are still in the prototype stage or just beyond it, IBM says adoption is poised to boom. A recent IBM survey shows that 56 per cent of Canadian respondents plan to increase AI investment in 2025. Looking five years ahead, AI adoption and implementation into the supply chain is expected to be five orders of magnitude larger than today, with many large firms expected to find effective use cases in the next 12-24 months.

Salley concurs. “Within 12–24 months, we’ll see proven AI applications in forecasting, inventory balancing and lead-time prediction,” she said. “But success depends on modular, composable systems that evolve over time. It’ll also require leaders who understand that even the best AI needs human partnership to be effective.” She added that within the next two years, AI will streamline forecasting, automate replenishment and predict disruptions. In five, “expect intelligent networks that self-optimize, which is adapting in real time across design, planning and operations,” Salley said. “But that future hinges on blending AI with experience, not replacing it.”

Meanwhile, AI in the supply chain is expected to follow the historic pathway of other technologies. “As AI companies gain experience deploying their tools, and clients become more familiar using them, the cost and effort required to deploy them are expected to decrease,” explained Haesen and Cascone. “At the same time, clients’ confidence and willingness to adopt AI are likely to grow as they see its effectiveness and the tangible benefits it delivers.”