steps to get started with generative AI in supply chains
With more companies beginning to leverage generative AI to optimize operations and enhance customer experience, it’s evident that the technology is here to grow. In our blog, Resilient supply chains using generative AI we discussed how generative AI is creating value for supply chains through multiple use-cases. In this edition, we cover some best practices businesses should consider as generative AI becomes a more permanent part of supply chain operations.
Identify specific use-cases
Begin with a focused, well-defined use case that addresses a specific challenge your supply chain faces today that is difficult to achieve through traditional ML approach. Start with a clear understanding of what you aim to achieve.
Build a supply chain data hub
The data hub would generate a single, harmonized view of supply chain data from various enterprise and execution systems, which will be leveraged by LLM models for real-time planning, simulations and forecasts.
Take measures for data privacy and security
Ensure robust data security and privacy measures are in place before deploying foundation models or generative AI solutions. Review terms and conditions to validate data privacy, deletion and storage mechanisms.
Determine the right model as per use-case objective and cost involved
Foundation models can be purchased or open source licensed to be used as-is, fine-tune or engineer prompts to improve pre-trained models for a specific outcome or even custom-build models for private use.
Collaborate with data and analytics experts
Collaborate with experts who have varied experience in deploying AI/generative AI solutions—specific to supply chain challenges and can customize as per your business needs. Ensure that the experts are involved in the implementation process to guide, test and fine-tune the models.
Start with quick win projects and track performance
Implement proof-of-concept pilot initiatives to drive quick wins, articulate and continuously track the incremental business value when striving for scalable adoption. Implement comprehensive mechanisms for continuous monitoring and verification to assess reliability, accuracy and performance.
Build CoE for scalability
Set up a generative AI Centre of Excellence (CoE) with a network of systems and processes, each focused on a different domain or line of business, and each with combined technology and business experts. These CoE ensure quick, easy to replicate, and accurate results with generative AI models for each instance.