The first thing I would do is slow down and identify the actual regulatory question. First establish intended use and risk. Search for a suitable predicate. If there is one, assess substantial equivalence. If none exists and risk is low to moderate, evaluate De Novo. Higher-risk products may require PMA. That gives you something you can check, discuss, and defend rather than an answer based on memory.
Here is a simple example. A novel diagnostic software product with no suitable predicate may be a De Novo candidate. A new version of an existing device might still follow 510(k), depending on its intended use and differences. That is why I would always ask for the intended use, the device description, and the relevant evidence before making a recommendation.
What does someone in RA actually do here? Often the junior person does not make the final decision. They research the official sources, compare possible routes, gather the right documents, flag gaps, and write a short rationale that a senior colleague can review. That is valuable work, even if you are not the final signatory.
Try this exercise: Prepare a pathway decision memo showing alternatives, evidence gaps, and questions to resolve with FDA. Keep it to one page. Add the source, date checked, assumptions, and unanswered questions. If your reasoning changes after checking the official material, that is learning—not failure.
One important limitation: A real FDA recommendation is case-specific; do not promise a 510(k) will work against contrary agency feedback. These decisions are specific to the product and market. A short internet post is a starting point, not a regulatory determination.
For people working in device RA: what is the first question you ask when someone brings you a new product or a proposed change? I think those practical questions are what newcomers most need to hear.