How can artificial intelligence improve supply chain risk management?

Consumers never buy the product, they buy the benefit out of the product. With IOT the world is so connected that people spit out data without knowing that they have been doing so for years now. AI could use those human behaviois to learn about human needs so that they can be cateres to. What we dreamt as impossible a few decade ago is today's reality. AI can learn from the past and suggest the modifications needed in future products this reducing cost and improving usage.

In addition to what has been already mentioned by other Linkedin friends, AI/Robots can practically become the Back office operations for SCM Contracts and Agreements analysis and reduce risks Help in Business Continuity management of org Integrating with Blockchain, it can improve regulatory, compliance and transparency/visibility AI can support SCM to increase both efficiency & effectiveness

Artificial intelligence can enhance supply chain risk management by providing advanced analytics and predictive capabilities. AI algorithms can analyze vast amounts of data to identify potential risks and trends, allowing for proactive risk mitigation. Machine learning models can improve demand forecasting and inventory management, reducing supply chain disruptions. AI-driven real-time monitoring and automation enable rapid response to emerging risks, enhancing resilience. Furthermore, AI can optimize supplier selection and performance tracking, fostering a more robust and agile supply chain that is better equipped to handle unforeseen challenges.

AI can support SCRM, but it's too early to use it to generate accurate data at scale. At present leveraging only AI to produce insights that can be used to make SCRM decisions, can represent a risk. However, an organization can use AI to summarise data, extract information, arrange data into a different view, but overall get a first high level feedback. Only validated ML algorithms that are specifically trained to perform a certain task, with large data points, can realistically be used to automate a insights, recommendations and predictions. Explore the various algorithms in the wild to define which one makes sense for you, as unfortunately just using Open AI API without defined data models to further train the algorithm won't be accurate.

AI can help buyers to pick right suppliers based on past data for performance and most importantly cost. Using AI effective cost negotiations can also be done based on different parameters including changing dynamics of market place, economic disruptions and other areas. Imagine a CHAT GPT type tool for buyers where they can query a large dataset using prompts based on their needs and requirements from their supply chain and getting faster effective results in seconds for which otherwise lot of research and analysis is done in current times. Off course we can’t fully rely on those results but surely that will help in reducing the turn around time for closing a better deal with good quality.

What I've done using AI was to create predictive capabilities. These are powered by data, monitoring and machine learning. For example, we are able to tell how much stress a certain function can withstand before breaking down, thus managing the resources required in the midst optimal way. Adding on to predictive capabilities, we can also use it to manage exception handling and business continuity

I think AI will give us some recommendations, but last word would be belongs to us. We can't give the complete control of system to it.

The application of AI capabilities on models like JIT (Just in time) by training the solution to identify the trend of orders and reorders to enable automated restocking without the likelihood of human errors while continuing to avoid the risk of overstocking and potentially tieing down working capital and other resources.

The greatest application of AI would be the reduction of the human interference in the supply chain flow thereby mitigating the risk of conflict of interest, internal fraud, misappropriation, bribery etc. Replacing humans with AI in the first line in the supply chain management and moving humans to oversight increases efficiency and reduces risk within the chain.

AI is definitely faster, more reliable and accurate than humans when it comes to manual repetitive tasks like monitoring, data analysis and recognizing patterns. However, I'll think of AI at this point as an optimizer to humans as certain critical decisions cannot be delivered by AI. Yet.

AI can be used to persuade us by logical explanations about stuations or problems. AI can give us examples or possible scenarios and their consequences may be good ways to persuade us.

The largest challenges and limitations for AI in SCRM are the validity of data it uses for analysis. We need to be able to trust the data that is being used, and potentially be able to review that data when we receive results that seem skewed. I fear too many will accept bad data and results as truth, and will not use critical thinking skills to determine the validity of the information being received.

AI-based SCRM is an ongoing process that necessitates human supervision to address potential system oversights. For example, in contingency situations generating product allocation constraints, where the system must decide where to ship medical products from a particular warehouse, it is essential to validate whether customers have stock or are out of stock. Relying solely on distance considerations for delivery may adversely impact the health of patients without medication. Human oversight is crucial to prevent such risks. Furthermore, implementing a robust reinforcement training plan for human-system interaction is imperative. Users must be vigilant for red flags and avoid complacency, which can be detrimental in abnormal situations.

In my previous point, I mentioned that AI can only be an optimizer at this point as certain critical decisions, especially if it involves moral, ethics social and other subjective topics. Humans should learn and trust AI so that AI can take off load them to focus on actually tasks that requires critical thinking

As I said we can't give the complete control of system to AI. AI can be a good adviser, but which should be verified. We can't one hundred percent trust AI.

Couple AI models with specialised resources is the way to go. Also before expanding into full AI, you should have a data strategy defined for your organisation and also define what components of your SCRM you want to automate, as realistically not all can be automated just yet, without running a second, human led, check.

I guess you can think of AI as a very smart and fast new employee so how would you onboard him? Define his job scope, what is the problem, and what is the objective. Inform him of the resources available, time, data etc that he can use. Draw up the parameters and boundaries, help him to not cross over to unnecessary terrains. Ensure that the resources and boundaries, especially data is always updated. If he still doesn't know what method to use, then tell him what's available. Observe, iterate and update accordingly.

IA can be a perfect teacher and advisor but can't be a decision maker. We can use it in some outomated processes like a perfect worker. But, it can't be a boss.