You can build based on intuition, but it's better to build based on knowledge
Assuming you have a solid understanding of your customer jobs to be done, pains and potential benefits, create product hypothesis and test them with your users.
Map it. Build it. Test it. Analyze. Repeat.
This can be fairly simple in B2C startup but very complex for B2B startup. Identify hidden jobs to be done which create huge opportunities for startups.
These might change over time with additional knowledge, but keeping track of them and communicating to them team is invaluable
You are building and MVP. How minimal can you keep it while making it viable? Change the the guessing game into a process.