For example, the intelligent application management system 430 of FIG. 4 may provide a query result 515 of a probabilistic model from a query “Can we decommission Db2 enterprise server edition on Hostname: ABC123 for account ACME?” for a software application. The key search terms (with aliases) may include “DB2 Enterprise Server Edition” and “ACME” and yield the query result 515. Query result 515 may be a table that identifies in the columns the search terms and the number of times each type of data (e.g., structured, semi-structured, and/or unstructured data) is references along with a combined total. For example, row 1 identifies the software application (DB2 enterprise server edition) as a non-potential software decommission candidate from the query with the search terms of ACME, ABC123, and DB2 enterprise server edition. The search result identified 5 structured data files/records, 10 semi-structured files/records, and 15 unstructured data files/records with only a total ranking score of 30.

In contrast, row 2 identifies the software application (DB2 enterprise server edition) as a potential software decommission candidate based on the query with the search terms of ACME, host: DEF123, and DB2 enterprise server edition. The search result identified only 2 structured data files/records, 0 semi-structured files/records, and 0 unstructured data files/records with only a total ranking score of 2. In one aspect, the ranking assigned a weighted value of “1” for each identified data source and added the each of the weighed values to yield a total ranking score. Other scoring methods and assigned values may be used according to user preference.