4 notes filed under numbers, newest first.
Most of it. Ramp leads with the buyer’s outcome, puts one action in the first screen and repeats it, links its big number to the method, asks a one-question quiz in the buyer’s words and answers the switching worry on a button. The live counters, the video and the 85,000 customers need Ramp’s scale.
For the decisions and the checking. AI drafts research, first passes and boilerplate code. I decide what to build and what to leave out, make every design decision, read every line before it ships, and answer for the result. The price is fixed for that result, so how fast the AI was is my risk.
Add up the hours your team spends re-typing and checking the same facts, then the errors that get caught late. If one year of that costs more than the tool, build it. If it doesn’t, or a product you can buy already fits, don’t.
Fast enough is Google’s ‘good’ line on three measures, for three visits in four on phones: the main content in 2.5 seconds, a reply to a tap within 200 milliseconds, and a page that barely jumps. Below that line, fix speed first. Above it, I would usually spend the money on the offer.
Two free tools go with these notes: a site check for people and for AI, and a payback sum.
Free tools