I'm not sure how you can avoid creating a function/procedure or template.

I suggest using INFORMATION_SCHEMA views to generate the dynamic SQL.

Here's some sample SQL, but you'll probably need to CAST the "newid()" or a set of case statements to match your random value to match the appropriate datatype.

SELECT

schemaname = c.TABLE_SCHEMA

,tablename = c.TABLE_NAME

,columnname = c.COLUMN_NAME

,c.DATA_TYPE

,c.CHARACTER_MAXIMUM_LENGTH

,sqlquery = 'select [' + c.COLUMN_NAME + '] = newid() from [' + c.TABLE_SCHEMA + '].[' + c.TABLE_NAME + ']'

FROM INFORMATION_SCHEMA.COLUMNS c

--WHERE c.TABLE_SCHEMA = @schema

--AND c.TABLE_NAME = @table