But in 1985, dBase and its clones were ascendant, and they were plentiful and programmable and somewhat language-compatible. Lots of people were employed hacking on dBase code for small businesses. (I also wonder if piracy made dBase more ubiquitous...)
Essentially Cornerstone. They made large investments in developing, marketing, and selling it and it basically flopped although it had a variety of innovative features from a technology perspective.
Arguably they were also slow to branch beyond pure text adventure games but the investments in business software were the bigger near-term problem.
Sounds like everything they built was really good. Read the piece, but still hard to piece together why they failed.