I work as a technical information secuity trainer (mostly) at KPMG Finland. We are very much a Powerpoint centric organization. I have maybe ~5000 unique pages of training materials that we have been produced during ~20 years. Updating, modularization, collaboration and versioning is a pain. Counting all the unmanaged copies and versions, we probably have ~1000 Powerpoint files with 50 000 slides.

I recently tried, as a proof-of-concept, preparing a course with Markdown -> Pandoc -> Powerpoint, and I am pretty happy. An unintended nice surprise was how easy it is to change Office themes with the output. The company (and different departments) want to change the slide deck themes every now and then, and customers might want to use their own themes on per-customer company tailored courses. 

My pipe dream is to start keeping all training material, including binary lab resource in Git, and building a perfect hierarchy of includes to easily produce different trainings from the same base material, which would be always updating. So that when sildes about, say, "Hardening TLS" would be updated, they would update to 10 different courses automatically. 

Next dream is to automatically produce tenders, CVs and what nots from the business databases? Write scripts that combine some Markdown templates with some ASCII reports and convert all to fancy marketing brochures? 

I also dream of adding Git versionable and scriptable graphs with dot and Graphviz.

Has anybody ever had similar dreams? ;-)