Generative artificial intelligence (AI), such as ChatGPT, has been making significant advancements in creative industries. However, its integration into the intricate world of supply chain management has been a slower process. Nonetheless, experts speculate that the synergy between generative AI and other AI-powered data analysis tools could potentially reduce the need for human intervention at various stages of the supply chain.
The question arises: Does generative AI hold genuine promise within the supply chain, or is it merely a hyped trend? The answer varies from one organization to another and hinges on a thorough comprehension of the challenges and possibilities associated with generative AI in supply chain operations.
AI applications have a longstanding presence in the supply chain domain. Machine learning and algorithms have been employed for years by supply chain professionals to enhance data analytics and aid in tasks like demand forecasting and decision-making. The appeal of generative AI lies in its ability to take insights derived from traditional AI applications and make them more accessible, searchable, easier to interpret, and beneficial for a broader audience.
However, existing generative AI models come with certain limitations that can hinder their implementation. Notably, the lack of transparency in how these models generate responses has raised concerns. OpenAI, the organization behind ChatGPT, has faced criticism for its lack of transparency regarding the data used in its product’s training.
From a compliance standpoint, this lack of transparency poses a significant challenge for supply chain businesses. For instance, achieving specific ISO certifications requires meticulous documentation of processes and adherence to numerous requirements. Overreliance on generative tools may lead to situations where businesses cannot explain the basis of their decisions, jeopardizing their certification prospects. Moreover, ChatGPT’s inability to access events post-2021 limits its effectiveness in adapting to socio-economic or environmental changes that could disrupt the current supply chain.
In summary, the current state of generative AI necessitates further refinement and definition before it becomes a reliable tool for decision-making.
It’s important to note that not all generative AI tools are the same. Some, such as Google’s Bard or IBM’s watsonx, have the capacity to access more recent online information as opposed to relying on outdated data. Additionally, emerging startups are advocating for an open-source approach to generative AI, which can address many of the visibility concerns raised by supply chain leaders. In essence, there is a path forward for generative AI in the supply chain, and its potential should not be overlooked, despite the current challenges.
To prepare for the adoption of emerging AI models, it is crucial to understand their limitations. Like any new technology, there is a risk of succumbing to hype and rushing into implementation without proper planning. By exercising patience, organizations can fully harness the benefits that generative AI offers.
Here are three key areas in which safe experimentation with generative AI is feasible:
- Administrative tasks: The supply chain involves a plethora of logistical tasks and document creation requirements. Experimenting with generative AI tools in these administrative areas carries minimal risk. For instance, these tools can be utilized to generate customs documents like packing lists and commercial invoices. Training generative AI on extensive datasets of such documents enables it to learn structures and recurring patterns, ultimately enhancing quality over time. This approach can be extended to other departments, such as human resources, customer service, and basic coding tasks, thereby streamlining routine activities and freeing up time for more valuable work.
- Demand forecasting: While traditional AI tools aid in demand forecasting, integrating generative AI offers increased flexibility and accessibility. Generative AI can incorporate data from a broader range of sources, reducing reliance on historical data for predictions. When using generative AI for demand forecasting, specificity is vital. Clear prompts that specify data sources and desired outcomes can help mitigate the risk of receiving vague or inaccurate responses, especially when compared to previous forecasts using traditional AI tools.
- Supply chain optimization: Generative AI with access to relevant supply chain data can be employed to optimize various aspects, from supplier evaluation to route planning and logistics. By aggregating data from financial reports, news updates, and other resources, organizations can build comprehensive profiles of potential suppliers. When it comes to route planning and logistics, generative AI algorithms are being trained to integrate real-time information, such as traffic and weather data, enabling organizations to optimize delivery routes and anticipate disruptions, ensuring efficient transportation of goods throughout the supply chain network.
Most supply chain businesses are increasingly recognizing the value of AI in their operations. Generative AI represents the next phase, offering new opportunities for real-time data analysis and streamlining administrative tasks. While integrating generative AI alongside other AI solutions will require experimentation and refinement, the result promises enhanced efficiency and insightful decision-making.
