I enjoyed your 2016/3 editorial ‘From linear programming in 1958 to winning at Go.’ But I find it hard to believe that you could discuss AI in oil and gas without mentioning important applications like Schlumberger’s DipmeterAdvisor and the use of constraint-based reasoning for automating the configuration of equipment like BOPs, equipment maintenance advisers and tools for well safety, inspection and regulatory compliance. These deserve attention because while there have been some ‘hitches,’ like DipmeterAdvisor in fact, they have, in many cases, been commercially successful.

I understand that those coming from the data side of the IT house may be more comfortable with the analytical, ‘computational intelligence’ approach with its roots in linear programming. But us AI ‘symbolists’ are still here and we are still making a difference, particularly as the industry hemorrhages domain expertise due to economics and the age of the folks with the heuristic knowledge.

It is especially important today to implement, test and deploy applications that capture and share knowledge across the enterprise. Analytical applications don’t do this, even though they can add value. It isn’t an either/or proposition, fuzzy logic/neural net hybrids are extremely powerful. You mentioned Google’s GO app which was a hybrid as is IBM’s Watson.