You are in a bind, my friend.
If those funding your work expect you to generate great design without any knowledge of users, well that's a fool's errand. Such organizations generally don't really care about great design. And so your work might represent somebody's idea of a checklist item for what it takes to build software.
While analytics could reveal something of value about current users, such numbers don't appear forthcoming. Drats! More valuable, anyway, would be data reflecting needs of people who will be using your product. But your hands are tied there because you aren't allowed to conduct interviews or focus groups. What a riddle!
Other sources of raw material you did not mention include casual conversations with people at the company (or a partner company) who might shed light on the context and needs of potential users. Perhaps of even greater value is haunting the Internet in your domain. Even where super-double-top-secret subjects are in the mix, it's often possible to find discussion groups hosted by Yahoo groups, Quora, and even on Facebook where people will lay it on the line.
You aren't clear whether the no persona rule is one laid down by you or by someone else. Still, it's true that without research data nobody has any business creating personas. Personas are difficult enough to compose and use with confidence... doing so without actually knowing your user base is irresponsible.
Still, I got to say, I've designed literally hundreds of products in my time, always with the help of personas and scenarios, and they never fail to help all involved understand what is at stake for the people on the other side of the glass. An organization opposed to basic research and synthesis methods must be challenged. If satisfying users is so important to their strategy, what better way to convey human needs than to show how humans (or models of humans such as personas) use your product? (Showing lines of code is not an acceptable answer.)