See links to support theory:

It appears that AGID can be run from blood, and is an indicator of exposure due to the presence of antibodies. It does not indicate specific types of influenza.

“Rapid diagnosis can be made by the detection of viral antigen in tissue impression smears using immunofluorescence, or by antigen detection enzyme-linked immunosorbent assay (ELISA) on tissue homogenates. Pancreas and kidney are the organs in which antigen is most often demonstrable.” (link)

This suggests AGID is easier/ faster and perhaps more cost efective for testing large numbers of specimins with further testing only on those testing positive.

The AGID test is an Agarose Gel Immunodiffusion assay. Basically, you put an antigen of interest (a virus) and a serum sample from an animal into adjoining wells of an agar plate. The antigen/antibody complexes that form (if any) should form a line between the two gels, basically solid precipitates in the agar. As far as I can tell, the specificity and sensitivity of this test (as you would expect, based on eye detection) is really low. ELISA analysis with an avian influenza virus antigen would be more accurate, if it is available.