I should have specified my threat model above, I'd ideally want zero-knowledge encryption on all tools, since Google is a threat in my model: assume that the NGO has a database of whistleblowers that needs protection