AI in Sourcing Optimization: The Day After Tomorrow (Part 1)

In the first article of this Spend Matters PRO series, we recounted the story of AI in optimization today, or, more accurately, the lack of artificial intelligence in optimization today. While AI in its most base form of "assisted intelligence" is readily available in many modern procurement and sourcing platforms, as evidenced in our previous two series on AI in Sourcing and Procurement, it has not yet creeped into optimization.

But that doesn't mean that AI will not creep into sourcing optimization tomorrow. While we may not see AI creep into any of the current platforms on the market (for different reasons for each vendor), that certainly doesn't mean that the next vendor to bring an optimization solution to the market won't learn from the oversights of their predecessors. In fact, in looking to get an edge over the existing, established platforms, it will assuredly be the case that tomorrow's optimization platforms will not only have a greater focus on UX and automation, but on AI.

Now, as per our last article, AI won't be embedded in the optimization engine — because that has to be powered by mathematically algorithms that have been rigorously proven to be sound (no errors, ever) and complete (will examine every solution and find the best one guaranteed) and given that most "AI" today (in the assisted and augmented intelligence category, because there is no true AI) runs on statistical or probabilistic machine learning algorithms, they just don't make the cut. (Although they can be paired with sound and complete MILP algorithms based on simplex or interior point to find faster starting solutions to difficult problems.)

But what about the day after tomorrow? What will we see then?