As far as i understand from your query, i believe there is nothing wrong in the way you are doing right now.
But you can always use SAP connectors and connect to SAP and fetch the data from SAP directly , instead of getting the data into excel, then loading them. But SAP connectors are not freeware, it comes with a cost.
So, if you want to eliminate the cost, then you can get the exports from SAP into excel and then use them.
With regards to cleansing and transformation of data, i believe you can completely depend on Qlikview and that the beauty of this tool. You can do it without setting up a data-ware house which is again an additional cost.
I hope i was able to answer few of your queries.
You have to take into account systems architecture, security issues, volume of data you are processing, data transformations you are doing. Maybe there is a reason that IT suggested extra database (for example aggregations on large datasets is better to do on SQL side not Qlik's). Or they just don't understand how Qlik works.
Who told you that you are supposed to use a Data warehouse? But having said this I would not go via Excel. But sometimes this is the only option available. And the administrators will not set up a direct Qlik connection.
But I try to avoid a separate non Qlik data warehouse as it involves another level of complexity and cost. And fortunately Qlik is powerful enough that a separate non Qlik DW is not needed (maybe with huge amounts of data but that's not my field).
But Qlik can sort of follow a data warehouse approach anyway.
I (unless its a simple one off App) extract data from say SAP (using Qlik) and save as QVD tables on an incremental extraction basis. I then load from these QVD tables. Some (and I have at times when heavy transforming of Data is required) even go a step further with a three step process. Extract data from SAP. Transform this data and save again as Stage two QVDs. And then load from these stage 2 QVDs into the final Apps.