There is no universally correct screen on which to ask for money. The useful starting point is what the customer understands at that moment and what service you are offering in exchange.
A paywall shown before any demonstration may be reasonable for an established, clearly described product. For an unfamiliar tool with an ambiguous outcome, it may ask someone to buy uncertainty.
Describe the boundary
List what is free, what requires payment and what happens when a subscription ends. Use language that matches the actual implementation. If an export is paid, explain that before a user invests an hour preparing data they cannot take out.
For a hypothetical research app, a free sample could demonstrate the evidence available while a paid plan funds a larger ongoing workflow. The sample should be real enough to assess quality, not a fabricated dashboard of perfect results.
Make the offer understandable
Show the price, billing interval, renewal terms and important limits. Distinguish a trial from a permanent free plan. If an annual price is displayed as a monthly equivalent, make the upfront charge clear.
A customer should be able to explain what they are buying without finding a tiny footnote or opening a support ticket.
Test the whole experience
Compare activation, paid conversion, refunds, support questions and retained use. A paywall that increases immediate purchases while producing more confusion may not improve the business.
Include cancellation and recovery states in the design. A failed payment should lead to an understandable path, not a loop of upgrade buttons. A cancellation should follow the terms the customer accepted.
Study competitor paywalls as examples of offers and information hierarchy. Do not assume that a visually aggressive screen is commercially successful because it exists.
The strongest paywall is a clear decision at the right point in a useful experience. Its job is to explain an exchange, not to win an argument with the visitor.