An office. A logo. A business card. None of them tells you whether anyone needs what you are building. In 2026, I would start with AI, content and community.
An office can be useful. A good identity can help. But neither answers the first question: who needs what you are making, and why would they care?
AI, content and community: begin with a problem
Start with a problem you understand. Use AI to explore a rough solution: organise your research, sketch a prototype or turn an idea into something a person can respond to. Treat the result as a first attempt, not proof that the idea works.
However, you still need judgement. Check the output, speak to real people and notice where your assumptions fall apart. The useful question is not how much you can generate. It is what you can learn from what you make.
Next, explain the problem in your own words. Share the attempt, the question or the part you have not solved. A post gives someone a chance to say: I recognise that. Or: you have misunderstood what matters.
At the same time, do not confuse attention with demand. A popular video and a useful product are different things. Look for conversations that help you understand a real need.
Turn feedback into the next attempt
Community begins with people who have a reason to come back. For example, respond to questions. Follow up on feedback. Let people see what changed because you listened.
Try one small cycle: choose a problem, make a rough version, explain it publicly and ask for specific feedback. Then improve it. AI, content and community become useful when they are connected to that learning process. The logo can come later.
For a concrete example, follow the challenge I set myself building Be The One with AI.
For example, finish a conversation by writing down one thing you had misunderstood. Then name the change you will try in response. On the next attempt, ask whether that change helped. This gives the feedback a visible consequence and keeps the next version connected to a person’s actual experience.
