Mid-infrared (FTIR) is the most popular tool for reaction monitoring because of the familiar chemical information that it provides. Changes in the ir spectrum over time can be interpreted as the changes in functional groups, the chemistry can be understood, reaction kinetics derived, side reactions and by-products can me minimised or eliminated. Diamond tipped ATR probes are used almost exclusively because of the chemical resistance of diamond and it'd reasonably broad infrared transmission window. Other ATR elements can be used where chemical compatibility can be checked and assured - these materials, such as Zinc Selenide will have higher throughput than diamond and cost less.
Raman spectroscopy gives similar information on functional groups to the mid-ir, with a rule of thumb that weak peaks in the infrared are strong in the Raman spectrum and vice versa. The other key difference is what each technique "sees": Infrared only samples materials in intimate contact with the ATR crystal, so it "sees" the solvent and materials in solution, but suspended solids are not measured. On the other hand Raman is a scattering technique and the denser the sample, the greater the scatter, so suspended solids will give the strongest signal alongside the solvent and dissolved solids. So whilst Raman is the equal of infrared for reaction monitoring, it is to be preferred for heterogeneous reactions where information on the solids is required and for crystallisations. Raman is not used so frequently mainly because of lack of familiarity with the technique.
Where the chemistry is understood and the requirement is to monitor progress and determine an end-point then UV-Visible can be very useful. ATR sampling can also be used with UV-Vis, using sapphire sensing elements in this case. The big advantage of UV-Vis is the significantly lower cost than the other two techniques.