The scenario “stuck at 80% completion” is likely.

Focus on the functions/features you currently use and can not live without.

Try to build an MVP (or as many microservices as needed) which serve as a proof that you can build the missing 20%.

Don’t bother building flashy GUIs, try to build the actual logic first. If you fail building this 20%, you avoid wasting funds for the flashy 80%.

Been there, done that. We built the flashy 80% and everything came to an abrupt stop when leaders realized we can not build the most important parts because the underlying platform does not support what was supposedly supported (miscomm with standard platform supplier during req engineering).