The box on the first screen is forgiving: you can start with almost anything, and you can stack extra context onto it to point the research where you want.
What you can start with
- A company or a website. A client, a competitor, or your own site. A specific inner page (a pricing or product page) is sharper than the homepage.
- An industry or a market. "dental practices", "law firms", "e-commerce brands".
- An idea. "an AI agent that books demos", described in a sentence.
- Your own product. What you're already building, in a line or two.
- A claim or a hunch. Something you already believe about the problem. The tool checks it against real evidence instead of taking it as given.
You don't have to pick just one. Mixing is where it gets good.
What you can add to sharpen it
Once you have a target, stack any of these onto it, in plain words:
- Your own take. "Linear, but I think non-engineering teams struggle with it." The tool tests your hunch against what people actually say.
- Who it's for. "Linear, for agencies." Re-lenses everything through that group.
- Compare it to a rival. "Linear vs Jira." Surfaces the gap between the two.
- A specific focus. "Linear's onboarding", or a specific page like "linear.app/pricing". Zooms the whole read onto that one thing.
- A market. "Linear in Israel." A local angle others miss.
These all work whether you started with a company, an industry, an idea, or your own product. The tool reads what you gave it, reflects back how it understood it, and points the research accordingly.
The more specific you are, the sharper the result. A precise page plus your own take beats a bare homepage every time.