A method, system and computer program product for receiving a spoken request to obtain indexed results from a database. Like result types are assigned to categories, and within each category is a plurality of result entries. The result indices are hexadecimal encoded, and each hexadecimal encoding is preceded by an initial character representing the result category. A speech recognition system is engaged, which processes the spoken request. When a item is requested, the respective category is implicitly known by the index returned, and the index provides direct access within a database to the corresponding result based on the phonetics of the request.