Whereas the Text Encoding Initiative (TEI) has become the reference standard for encoding textual material of all kinds in the humanities, the power of the underlying TEI modelling infrastructure to deal with professional document management scenarios or even non-TEI based vocabularies could deserve more attention. The aim of my presentation will be to show concrete projects where I have contributed to use the ODD (One document does it all) specification language of the TEI to deal with such applications as the management of patent documents, the modelling of lexical resources or the integration of heterogeneous archival descriptions in the EAD (Encoded Archival Description) standard. Starting from an introduction of the TEI as a standard, I will try to conclude on its potential bright future as a real infrastructure for the humanities.