Inventory Management in the age of generative AI
A practical overview of how best to approach inventory management and how to choose the right technology to manage it
I’ve been thinking about how so much content out there talks about managing inventory efficiently but hardly anyone gives some practical advice in understanding managing inventory and different kinds of systems and processes to put in place to manage them.
I’ve created this guide, to give a practical overview of how to best approach inventory management and how to choose the right technology to manage it.
Disclaimer: This post covers physical inventory. It does not cover digital inventory. So, suitable for supply chain, ecommerce, logistics and field service teams.
The supply chain and inventory management sectors are increasingly embracing advanced technologies like AI, IoT, and low-code platforms to enhance efficiency, predictability, and resilience. These innovations are poised to significantly impact inventory management practices by improving data quality, enabling agile responses to market changes, and facilitating a more integrated view of supply chain operations. Source: KPMG
Let’s start with the basics
Simply put, it’s the process of managing the goods needed to run your business. You need to manage these goods at the right quantity, quality and cost to run your business efficiently. That’s all it is.
When you’re a small startup, you can order these items from a local seller, put them in your home and start selling them. But when you start scaling, you need to put the right processes in place. Because, managing inventory will make or break your business.
Managing inventory at the right levels will make or break your business. Here’s the general principle.
If you have low inventory, it will lead to stockouts. You don’t have products to sell and you will lose customers.
If you have too much inventory, you are going to block your working capital and increase your storage costs. Which add up to quite a lot.
Let’s taken an example here. We’ve considered different holding inventory levels for a store. All other assumptions are kept constant and we’re only modifying the holding inventory levels to figure out how it affects profitability.
Now, this is how our profitability changes depending on how much inventory you want to hold.
As you can see, your profit gets significantly impacted when the inventory held changes. Now, including variabilities in seasonality, customer interests and market factors, managing the right amount of inventory becomes a science. This aspect is called Economic Order Quantity.
Now that we have setup the importance of inventory management, let’s understand the objective. When you think about any process, it’s important to condense the objective to 2-3 items so it’s easier to process and manage. Inventory management function, typically needs to optimize for the following.
% stockouts - This is the number of orders that cannot be fulfilled on time because of low inventory. You’d want to maintain this as low as possible.
Inventory holding cost - The cost of storing excess inventory. This should also be as low as possible.
Once you understand these, it becomes easier to decide what is needed in an IMS solution.
Before we jump into what to look for in an IMS solution, it’s good to understand different types of inventory. Inventory can be largely bucketed into 4 types.
Raw materials are products you use to manufacture the items you want to sell. Example, cocoa powder and sugar become raw material to manufacture chocolates.
These are items that can be sold. This can be manufactured by you or can be purchased directly from suppliers. Example, if you want to sell shoes online, you might purchase it from suppliers instead of making them yourselves. Or you can store the chocolates manufactured by you in your inventory.
Consumption items are products that you do not sell or manufacture but you need to run your operations. Example, you might use oil and grease in your maintenance facilities.
Assets are equipments you buy to run your operations. These tend to be machines to manufacture, pack, fulfill etc. Could also be laptops for people to work.
In this post, we will mainly focus on Raw material and sale items.
While we are here, I want to take a tangent and talk about a related topic because this impacts how you manage your inventory.
Made to stock items are defined SKUs (Stock Keeping Units) which are manufactured, stocked and sold to customers. There’s no difference between one item and the other, so the items can be manufactured ahead of time depending on expected customer orders.
For made to stock, you take the sale projections, identify inventory needed and initiate the purchase on time so you can start fulfilling as and when the customer places an order. Most ecommerce is done this way.
Made to order items are typically unique for each customer.
The customer places an order → Identify the SKUs needed to fulfil the order → Purchase those products → Fulfill.
A great example of this is Rolls-Royce cars. Each car is manufactured according to the specification of the customer.
Given what we know, here’s what a good inventory management system should do.
Procurement planning should solve for when and what items to order from your suppliers.
This system will use data like
The alerting system should indicate what items need your attention. A simple definition would be to identify items that break a pattern. Either they are getting over too fast or moving too slow, these need to be brought to your team attention so they can take action
Your IMS system should be able to given you 100% visibility into every single movement of goods, current quantity of goods available at each store etc.
Your IMS should also be able to manage multiple stores and multiple store types
Types of inventory stores
and more.
Generative AI has a big part in managing inventory. A concise way to put it would be that generative AI is transforming inventory to be “managed by exceptions”.
You no longer have to monitor your inventory. Your IMS with AI will notify you of any exceptions, ask for approval and perform the necessary task to resolve the exception.
For example, when a particular stock is running out, generative AI will be able to create a Purchase Order (PO) draft, notify you over slack for approval and on approval, can send it to your preferred vendor.
The ops world today works as Data → Analysis → Reports → Insights
With the advent of AI smarts, we can cut down the cycle and directly head from Data to Insights without the need of Analysis & Reporting.
Managing inventory the right way, imo, is the biggest factor in making a business profitable. In order to do this right, a lot of factors need to fall in place. This requires great discipline and iterative process. Hopefully, this is a good primer for you to understand how inventory should work for your organization and how you can choose the right product that suits your needs. If you have more questions around this, feel free to write to us.
Further explore our blog on the benefits and trends in quick commerce here.
