A common understanding of business documents is important for realizing interoperability among collaborating enterprises. In this paper we report on the design and evaluation of a system that can help collaborating enterprises to efficiently recognise and align business documents (e.g. incoming emails and attachments) according to the enterprises' local document repository. The system relies on the interplay between automatic recognition of business documents and human intervention by means of an assistive mapping tool. Our findings show that the balance between automatic recognition and human intervention ensures greater levels of accuracy in the system. We evaluate our system from both usability and accuracy perspectives and argue that some of the lessons learned and design decisions we took are applicable to the general mixed initiative tools for end user support. © 2013 Elsevier B.V.