How do you see the future changing for the XBRL process?

The future of XBRL(and its past) is rolling out the same way many other innovations or statutory requirements roll out.

First methods are implemented in a similar manner to how they were done in the past. This translates in the case of XBRL to many companies using their printers to create the XBRL mapping/tagging and instance documents.(my informal statistics show 85% - 95% of companies used their printers for the first year of XBRL filing).

Then in a second wave of change, the process and new capabilities are thought about in a different and maybe novel way, the original problem is framed in a different manner and innovation occurs.

In the case of innovation with XBRL, that change is occurring. While my informal statistics for first year tagging are 85% - 95% of companies are going to their printers for year one tagging (block tagging), in year two only 75% - 85% are going to their printers for detail tagging which is harder, and in year 3, only 45% - 50% are using printers and those that are using printers are using them in a much more diminished capacity.

Software is supporting this shift. I break XBRL software solutions into three levels.
1) XML Editors made to look like XBRL editors. These are characterized by interfaces requiring you to edit/dissect the XBRL US schema into your own schema, then map/tag against a text, word or excel document, then edit the schema for extensions or things you forgot then map and tag again and iterate through the process until you get it right. Very counter intuitive process.
2) Excel and Word client side tagging tools that work as an addin to Excel or Word. Still a very manual and cumbersome process but you are working from the final document interface, rather than the tool interface. You still have to make sure each number is tagged with the right context, role, parent (presentation), calculation, line item, axis, etc). Extensions are a three step process of 1) creating the line item, 2) setting the presentation (i.e. anchoring the new tag), 3) setting the calculation link base
3) SaaS solutions that have imbedded intelligence either via "regular expression processing" or "natural language processing" to automate 60% to 80% of the tagging process. These tools use as a reference the financial statements already files with the SEC by other companies for pattern matching so you can know how well used (or not used) a tag is. Inherently they have to be SaaS because their database is continually being updated with latest filings. The three software providers in this category are BCL Technologies, NeoClarus and Webfilings. BCL and NeoClarus use MicroSoft Office documents, Webfilings use google docs.

There is another innovation that is occurring to help public companies compile the financial statements (and related notes, etc) that are Edgarized to the SEC. These solutions componentized the compilation process so multiple people can work and collaborate/approve the SEC document independently, with workflow control, then the software creates the final SEC document and Edgarizes it. Those solutions are provided by (IBM, Host Analytics, and Webfilings). IBM partners with BCL Technologies to provide the Edgarization process. Host Analytics also partners with BCL Technologies to provide XBRL and Edgarization.