Korva is a scroll of the internet’s best ideas.
Measuring trust
Trust in an AI agent builds on a schedule. Instinct tracks how long a new user takes to share a first credit card, a first account password, or a first sensitive piece of information, and treats those times as proxies for trust. The 40 percent figure counts the whole user base, including people who left before that point.
Why access matters
The reason the number matters is that more access makes the assistant more useful. The more data it has, the more proactive it can be and the better it fits a person's situation, so each step of sharing raises what it can do for them. Users set the pace, and they can take any access back.
The retention payoff
The payoff shows up in retention. Among users who connect at least one piece of sensitive information, about 80 percent keep using the product.
Why do people share access in small steps?
Agreeing to a small request makes people more willing to accept a bigger one later. Psychologists Jonathan Freedman and Scott Fraser tested this by asking homeowners to display a small sign, and those homeowners were far more likely than others to later agree to a large sign on their lawn. The pattern is known as the foot-in-the-door effect. Their study appeared in the Journal of Personality and Social Psychology in 1966.
Why measure trust by what users do?
Stated trust and behavior often differ, since people say they are comfortable far more readily than they hand over a credential. Economists call the idea of inferring preferences from choices revealed preference. Behavioral measures such as the time to a first shared password can also be tracked automatically for every user, which surveys cannot do. The concept of revealed preference was introduced by the economist Paul Samuelson in 1938.
More from Invest Like the Best with Patrick O'Shaughnessy
