No need to require all partners to switch to the same system first. The article suggests adding a data transformation layer between different documents and systems: read the other party's files, extract content, then organize it into the fields the receiving end needs. APIs work well when partners already have system interfaces; document recognition handles cases where file exchange is still the norm. When implementing, confirm the recognition scope with actual formats you receive, check results column by column. Don't assume that 'reading any format' is a guaranteed out-of-the-box capability.