The pandemic disrupted supply chains all over the world and many enterprises turned to AI for help -- possibly in a way that doesn't make strategic sense.
As business and supply chain leaders are able to get some breathing room and examine their actions in reaction to the pandemic, many have concerns that AI adoption in the supply chain has moved too quickly. Now is the time to get strategic.
Here's a look at why that's important and what that means in terms of the future of AI in the supply chain.
COVID-19 hastened AI adoption
Many leaders have turned to new supply chain strategies in the last year, including AI.
Supply chains across industries previously relied on traditional models, said Sreekar Krishna, head of data engineering at KPMG. These models use statistical and time-series methods to create data that can be used to predict supply needs or forecast demand. However, few people could have predicted the pandemic. COVID-19's upheavals made these models less successful and revealed a new need for supply chain agility.
As the supply chain environment rapidly changes, leaders must exhibit a new level of adaptiveness and attention to detail, Krishna said. AI adoption is imperative because the rate of change in market dynamics requires organizations to react with speed and accuracy. Companies will need AI technologies in order to make significant, scalable productivity gains and stay competitive.
AI technologies can inform predictive models.
AI also helps organizations predict unexpected events and be proactive, said Bhava Kompala, global business head of retail at LatentView Analytics, an analytics consultancy in Princeton, N.J. AI's predictive capabilities can help companies build supply chain resilience so they can manage unexpected events more successfully.
Better guardrails for AI adoption
Some higher-ups feel AI regulation is lagging behind the breakneck speed of adoption.
"While not all industries are affected by regulations equally, some of the industry leaders are looking to regulations to provide guardrails to tamper the excitement within organizations [about] adopt[ing] AI capabilities systematically," Krishna said. Some are concerned that without these restrictions, AI will move faster than their industry can safely handle.
Part of the problem is that the pandemic is accelerating the pace of change, Krishna said. The level of retail disruption over the last year is equal to about five years of normal disruption.
Retail business leaders are among those who feel that AI adoption is moving too fast, according to the study "Thriving in an AI World" by KPMG, located in Amstelveen, Netherlands.
Their response is unsurprising, considering retailers are digitally adapting more quickly than planned, Krishna said. They're also navigating rapidly changing customer needs, which have affected supply chains and demand forecasting.
AI growth accompanied by unease over adoption speed
AI is at least moderately functional in a range of industries, according to KPMG's study on AI.
The study found significant variations between industries' AI development. AI is moderately to fully functional in 93% of industrial manufacturing organizations, 84% of financial services, 83% of tech organizations and 81% of retail, according to the KPMG report. Meanwhile, 77% of life sciences organizations, 67% of healthcare organizations and 61% of government organizations have moderately to fully functional AI.
Industries that have seen the biggest AI growth over the last year include financial services with a 37 percentage point increase, the retail sector with a 29 percentage point increase and the tech sector with a 20 percentage point increase.
However, some higher-ups have trepidation about the speed of AI adoption.
These concerns were most often expressed by industrial manufacturing leaders, with 55% mentioning the issue, followed by 49% of both retail and tech leaders.
Meanwhile, many AI leaders believe government regulations could sort out some of these concerns, and 92% of AI experts said government should be involved in regulating AI.
Other concerns about AI adoption could stem from these factors, Krishna said:
- the rapid progression of AI technologies offered by the market;
- lack of organizational readiness for digital transformation;
- lack of skilled resources, especially those who understand technology and industry; and
- declining investments in traditional companies compared to venture capitalist-driven AI investment in startups.
Educating on AI in the supply chain
Another AI adoption issue is a perceived overreliance on isolated tests that don't aim for true applicability.
Many supply chain leaders are focusing on pilots rather than engineering for scalability, said Suketu Gandhi, partner and global product leader of plan and digital supply chain at A.T. Kearney, a strategy and management consulting firm in Chicago. Many companies test the waters with AI, then end up with multiple pilot projects that don't provide a decent ROI.
"AI is showing up everywhere except the bottom and top line," Gandhi said.
Companies will have greater success if they develop a center of excellence (CoE) to understand how AI can truly help the organization, he said. That process should begin with educating business leaders about AI's capabilities, focusing on supply chain improvement.
"There is a missing translator piece between AI gurus and business folks that has been slowing adoption," Gandhi said.
Executives also need to understand how AI applications can create problems, including data quality issues that reduce accuracy. Ethics, governance and regulation surrounding AI and how those may create risks for companies are also critical concerns to address.
Sreekar KrishnaKPMG head of data engineering
AI in the supply chain needs work
Many business leaders are concerned about the pace of AI adoption, and will need to continue refining AI strategy through an agile approach.
"Halting any progress toward establishing the foundations for a successful AI implementation can result in your organization falling behind," Krishna said.
Project teams can work on data readiness, digital tool integration and skill alignment in parallel with developing an AI strategy and roadmap.
Organizations also need to execute the AI digital strategy as part of their product delivery lifecycle, Krishna said. AI capabilities can then become part of an agile process that allows for quick experimentation and rolling back updates if problems arise.
Retail and other supply chain industries should also continue developing foundational data capabilities, Krishna said. Data cloud migration, data quality initiatives, and business and supply chain optimization can all make AI implementations easier.
Companies likely won't see immediate positive results from AI development.
"Leaders must keep in mind that AI services do not necessarily create value overnight, and the investment is for the long-term benefits," Krishna said.