@carrucio
This does not match the requirement because "-e" just means "OR" and it will output all records which contain any one or more of the the three strings.

@poweroflinux
Though I still think that this needs to be a database program and based on your limited and inconsistent data sample, you can reduce the processing by using a workfile containing only the record type "CustomCDRInterceptor".
Also you need to "while read" $output in case there is more than one hit.
The whole script is dubious because of the chance of false hits.
I've left the "perl" alone because I don't know what it does and why the "awk" format is not enough (... unless it was dealing with multiple records in the same environment variable?). We really need to see representative data and required output and to know whether the order of the output is important.

Code:

grep "CustomCDRInterceptor" $2 > /tmp/workfile
awk -F"," '{print $1 , $2}' /tmp/workfile | \
while read a b
do
  grep "$a" /tmp/workfile | grep "$b" | cut -d"|" -f6 | awk -F"," '{print $4,",",$28,",",$27,",",$17,",",$12,","$21,",",$11,",",$26,",",$14,",",$6,",",$30,",",$31,",",$19,",",$5,",",$22,",",$10,",",$9,",",$20,",",$15,",",$29,",",substr($32,1,match($32,/\]/)-1),",",$23,",",$18,",",$24,",",$7,",",$13,",",$2,",",$25,",",$16,",",$8,",",$1,",",$3,","}' | while read output
   do
        #echo $output
        echo "$output" | perl -F, -lane 's/^\s*[- \w\[]+:(.*?)\s*$/$1/ foreach @F; print join ",", @F'
done

Depending on how big a workfile containing only "CustomCDRInterceptor" records becomes, there is potential to create a stuctured temporary file containing only relevant record types and relevant fields. Then to sort both this file and the search pattern file to a common order prior to running a single-pass match against 3500 records rather than running 3500 searches.