I loved it. I think the team lead managed it well. It helped that we already had insight into the old problem domain, and how it was changing. And experienced engineers can generally collaborate effectively, as well as work self-directed.
It's not that different than any project with a novel component. Some projects ("Uber, for cats!") are mostly a matter of executing in some well-understood ways as soon as you hear the idea (set up your tools, use these frameworks, on this cloud configuration, and have frontend and backend start grinding through the tasks, while you're plugging them into the plan, and scheduling for demos or MVP launch or whatever). But other projects have more upfront domain knowledge to acquire, brainstorming, requirements analysis, market analysis, prototyping, figuring out possibilities, deciding what you will do, etc. In the latter, eventually it probably turns into a conventional fairly predictable project (although there might still be tricky problems to solve in the pieces, but you have a good idea where you want to go and what the pieces are).
It's not that different than any project with a novel component. Some projects ("Uber, for cats!") are mostly a matter of executing in some well-understood ways as soon as you hear the idea (set up your tools, use these frameworks, on this cloud configuration, and have frontend and backend start grinding through the tasks, while you're plugging them into the plan, and scheduling for demos or MVP launch or whatever). But other projects have more upfront domain knowledge to acquire, brainstorming, requirements analysis, market analysis, prototyping, figuring out possibilities, deciding what you will do, etc. In the latter, eventually it probably turns into a conventional fairly predictable project (although there might still be tricky problems to solve in the pieces, but you have a good idea where you want to go and what the pieces are).