How do you decide which features to include in your product and how do you prioritise?

It depends on user feedback, but if you don't have any of that, then I would look for general trends. For example, I look at what my competitors are doing, then look at how what their doing solves a customer pain point, and what holes it leaves. So the same way I use to decide to create a product in the first place. Using my own startup as an example, there's been a lot of interest with ChatGPT and GPT-3 recently. However, these AI might be expansive, but they are highly restrictive in terms of what content you're allowed to generate, and quite costly as well. Competitors to GPT-3 will usually offer the inferior GPT-J or 2, which, although less limited and possibly cheaper, incurs a large decrease in quality. So what I'm working on is a happy medium between the two. Nothing can really beat GPT-3 in terms of quality, but I can definitely beat GPT-J or 2 but using a larger model (BLOOM 176B). The weak spots (lower quality) will be covered up by an addition to allow users to train their own models (fine tuning) in the near future. So using this process, I've already addressed two branches of my competition. If you're curious to see more about my startup, feel free to join my discord or follow my twitter in my bio for updates