In case you have not noticed, we live in the era of cheap enterprise technology. As an end-user enterprise, if you have the technical expertise, you can exploit this cheap technology readily. No matter how fancy or hyped the technology is. Back when ChatGPT was viral, there was news regarding Stanford engineers being able to replicate GPT technology with an investment of $600. Recently, I saw perspectives being floated that these replications are not as efficient as the originals. They probably are not, but do they have to be?

As an enterprise, you do not have to develop a know-all solution like ChatGPT. You need to leverage technology to develop solutions around your unique problems. So the gist is, from this point onwards, no one can control or dominate core technology, no matter how much they invest in developing that innovative technology. They can control solutions and innovations generated by leveraging those core technologies through IPs. But from an enterprise perspective, this is excellent news for you. If you can develop internal expertise, you can integrate any such advancement inexpensively.

This era of cheap technology means you can incorporate a portfolio of analytics solutions today that rarely any off-the-shelf solution provides. One such area is integrating recommendation systems in the world of supply chains.

We talk about automating supply chain planning. As I have indicated before, a hundred percent automation of supply chain planning is very far away. We can automate a good percentage of decision-making in the short and mid-term. Planners today rely on data and analytics to hone their intuition of decision-making. Augmented analytics promises to help them with this task by providing insights.

But ask any end-user honestly for feedback on any augmented analytics features they have used. Most of these augmented analytics features essentially regurgitate what traditional descriptive analytics BI solutions share. A box is checked in the solution from the perspective of the availability of augmented analytics. But augmented analytics’s real purpose was to provide insights that can actually aid in planning.

And if end-users want, they can infuse their forms of augmented analytics, leveraging inexpensive technology tools, in their current planning environments. What planners would love are insights beyond descriptive.

Recommendation systems are widely used, and we have interacted with them in our day-to-day life, knowingly or unknowingly. Recommendation systems are widely used, From products in eCommerce portals to movie recommendations in streaming portals. But not so much in the area of the supply chain.

One of the problems in supply chain analytics and planning is our fixation with traditional analytics approaches. So when a new technology comes on the horizon that can be leveraged for analytics, we try to leverage it to enhance existing methods. Now may be the time to infuse new approaches.

Consider an example specific to using recommendation systems in tactical supply chain planning. At the beginning of a shift in a DC, you have a pool of pickers and a pool of orders that need to be picked. Based on historical data, you can design a recommendation system to “recommend” the orders best suited for specific pickers. We have tried to solve these problems through optimization algorithms. But that approach, while better than randomness, is not the best.

And as you can imagine from the example scenario above, such tactical planning decisions abound in the world of tactical supply chain operations. Here is an example from transportation. If you have a diverse fleet with various transportation assets, we use our understanding to determine which assets need to be assigned to specific loads. Even in route optimization software, these constraints are inputs we provide. But what if a recommendation system learns from the historical data and recommends the best fit based on the characteristics of the route and load?

None of this is futuristic. You do need technical expertise, internal or external. There will not be a better time to jump on the technology explosion bandwagon.