prepare a 1M phosphate solution, and dilute 1:100 .Then dilute 1:100 again, 10ul into 1ml. Then dilute 2fold several times to make a standard curve. Run your assay, and then remove an aliquot and add your Malachite green components. Plot the standard curve, absorbance units vs phosphate concentration. Use the equation from the slope of the straight line to calculate the unknown phosphate concentration in your reaction.

I would do one thing differently from Marcia Moss's suggestion: I would use an arithmetic series of phosphate concentrations for the standard curve, rather than 2-fold serial dilutions. For example, the Malachite Green reagent I have used gives a response to inorganic phosphate in the range 0-20 µM, so I would make a standard curve of 0, 2, 4, 6, ..., 18, 20 µM. The standard curve need not be linear, by the way. It can be fit by nonlinear regression to any suitable mathematical function, such as a hyperbola or parabola.

The standard curve should be prepared in the same buffer as the enzyme reaction.

I agree with Dr.Shapiro in his answer