I would not focus on trying to replace SAP (which, to be honest, is nearly impossible for a startup) but to find ways to improve it or profit from its ecosystem

I would not focus on trying to replace SAP (which, to be honest, is nearly impossible for a startup) but to find ways to improve it or profit from its ecosystem. For example, Celonis is a German startup with a >1BN€ valuation that started by mining business processes from SAP systems to tell their customers how they could improve their practices and save money. SAP could've done this themselves of course but they ignored this market entirely, giving Celonis a chance to grab it (now they're partnering with them). I'm sure there are many other areas where you can create tremendous business value from the data in SAP, which SAP doesn't pursue due to them being too "niche" for them or due to ignorance. So I'd say find one of these cases and build an MVP around it, which is more realistic than writing a full-blown ERP anyway given your probably constrained resources.