It is well understood that an ensemble of problem solvers works better than any single of them, as a fact that is studied with Bagging, Boosting and Hierarchical Mixture of Experts. We bring this power of ensembles to a whole new front by employing autonomous problem solver agents who engage in problem solving organizational assemblies, and propose it as a superior framework for these sorts of problems. The organization has higher level nodes, namely, agencies, who keep track of other agents, but do not manage them. Experts have absolute freedom to choose their parts in the problem at hand, giving the assembly certain advantages. Implementation of this idea demonstrates considerable improvement both in terms of accuracy and speed.