The eternal AI quest - can machines think as well as man? - seems quaint today compared to the question of how can machines help man to think. True, Deep Blue can beat the world's best chess player, not by thinking, but by exhaustively examining all permutations and combinations in blinding time against a predetermined outcome set of rules. The questions for mankind, though, seem of the form where rules are imprecise at best, and essentially unknowable perhaps. If learnable and knowable even, many other constraints exist that mitigate against "data-driven decision-making". This presentation assesses some of these constraints, and offers some perspective on the value of using visual dynamics and analytics to help overcome such issues.