Your collaborators have their own world. They were already doing their work before you engaged them. They have their own tools, their own environments. If you're planning to force someone to use something else, are they actually going to use it?

Your platform could be beautiful with lots of capabilities, but if it's not easy for them to use and to get into, they're only going to publish into it, they're not going to work in it.

It could be fine for them to keep using their preferred workflow, but is the information going to flow? Will you get rapid support and enablement, or do you wait weeks to get access? Can you automate pulling the information down from their system into your system?

You have sensitive unpublished work outside this collaboration. What are your security controls? How do you segregate information? Can you segregate functional-level controls and not just data-level controls? Are there different levels of users?

About your collaborator's environment: does the architecture of this other thing fit into how you work, or how you want to work?

Do you have the basic operating model between the collaboration teams on both sides so that you can actually make the work happen and the data flow?

Every collaboration ends. When all is said and done, how do you get the information back out again? What's your back-out or decommissioning strategy?

You're promoting FAIR where you are - making data findable, accessible, interoperable, and reusable. But the people that generated the data don't even know that someone will want that data downstream, so they're not going to anticipate those needs. How can you get the value of FAIR and enrich the workflows of individuals with smart tools so that they see the value, too?