SAP’s Intelligent Enterprise Might Fare Better Without AI

IT companies (including SAP) like to talk about artificial intelligence (AI). They use modern buzzwords to bask in the short-lived glory of new trends. Blockchain and AI have become fan favorites, but sustainability is on the rise as well.

SAP wants to unite all the buzzwords under one common umbrella term: The Intelligent Enterprise. The ERP company wants to completely overhaul traditional supply chain logistics, making processes more intelligent and operations more – you guessed it – sustainable. It’s SAP’s vision, but without concrete use cases or success stories, it might never be more than that.

Artificial intelligence is not the only way SAP could go about making the Intelligent Enterprise a reality. There’s countless example of people doing great things without the help of AI technology.

Creating algorithms – without AI

For example, Jakub Tarnawski’s dissertation “New Graph Algorithms via Polyhedral Techniques” made groundbreaking algorithmic progress on two of the most central problems in combinatorial optimization: the matching problem and the traveling salesman problem.

Work on deterministic parallel algorithms for the matching problem is motivated by one of the unsolved mysteries in computer science: Does randomness help in speeding up algorithms? Tarnawski’s dissertation makes significant progress on this question by almost completely derandomizing a three-decade-old randomized parallel matching algorithm by Ketan Mulmuley, Umesh Vaziriani, and Vijay Vazirani.

The second major result of Tarnawski’s dissertation relates to the traveling salesman problem: find the shortest tour of n given cities. Already in 1956, George Dantzig and others used a linear program to solve a special instance of the problem. Since then, the strength of their linear program has become one of the main open problems in combinatorial optimization. Tarnawski’s dissertation resolves this question asymptotically and gives the first constant-factor approximation algorithm for the asymmetric traveling salesman problem.

Creating new algorithms or enhancing existing ones without AI is something the SAP community could benefit from. It’s time to refocus on real innovation instead of buzzwords. Which one SAP’s vision of the Intelligent Enterprise will be remains to be seen.