Predictions of working memory model?

If both tasks same component e.g. phonological cannot be performed successfully if 2 tasks use different components can work together somewhat well e.g. listening & talking