The instinct is to start with the largest problem, because that is where the value is. It is also where the politics are, where the requirements are contested, and where failure is visible.
What should the first internal tool solve?
Start with something small that annoys everybody equally. Where the logo files live. Which meeting room is free. Who has the current version of the deck.
Two things happen. You learn how the organisation actually works, in a place where being wrong costs nothing. And people who were sceptical start asking whether the next thing could work the same way. That second effect is worth more than the tool.