@aymnart fraud is one of our products and one we care very much about
Report
Love the idea of letting agents spend autonomously. One thing that would make me trust it faster is a built-in merchant whitelist so each agent can only charge approved vendors, with a clear audit log I can scrub through when the bill arrives.
The single use model does limit blast radius, they don't stop a compromised agent from spending that budget at the wrong place entirely. Curious whether there's any MCC level filtering on card issuance.
the single-use + destroyed-after-first-use design answers the double-charge question well, but I'm curious about the other side: disputes. with a normal card, "I didn't authorize this" has a human cardholder behind it that the bank's fraud process is built around. here the "cardholder" is a task an agent ran once and the card no longer exists by the time anyone notices something's wrong. if a merchant double-bills or a charge looks off after the fact, who actually files the dispute and does the issuer's chargeback process even have a lane for that, or does it end up as a support ticket with Agentcard instead of a real card-network dispute?
This should also be available for some influencers if they have an agent. Finally, some bots would use money more responsibly compared to those creators. 😅 [know one influencer who is wasting money literally for anything]
Report
One thing that would be super useful is a per-transaction approval prompt sent back to the human owner through chat, so I can quickly allow or deny a purchase without digging through a dashboard.
Report
Honestly kind of wild seeing my agent check out on its own without me babysitting the checkout flow. The spend limits give me peace of mind though, would be a nightmare otherwise.
Report
finally spun up a quick test agent and pointed it at a dummy checkout. it actually paused for confirmation before charging, which i was not expecting from a card pitched at bots.
Replies
How do you plan to handle fraud and liability with these debit cards, given that AI agents can potentially be compromised or act unpredictably?
Love the idea of letting agents spend autonomously. One thing that would make me trust it faster is a built-in merchant whitelist so each agent can only charge approved vendors, with a clear audit log I can scrub through when the bill arrives.
The single use model does limit blast radius, they don't stop a compromised agent from spending that budget at the wrong place entirely. Curious whether there's any MCC level filtering on card issuance.
Cool concept. Does each agent task get its own card, or do they all draw from one shared wallet?
love how focused the launch page is. the spec sheet showing spending limits per agent is such a practical touch.
the single-use + destroyed-after-first-use design answers the double-charge question well, but I'm curious about the other side: disputes. with a normal card, "I didn't authorize this" has a human cardholder behind it that the bank's fraud process is built around. here the "cardholder" is a task an agent ran once and the card no longer exists by the time anyone notices something's wrong. if a merchant double-bills or a charge looks off after the fact, who actually files the dispute and does the issuer's chargeback process even have a lane for that, or does it end up as a support ticket with Agentcard instead of a real card-network dispute?
This should also be available for some influencers if they have an agent. Finally, some bots would use money more responsibly compared to those creators. 😅 [know one influencer who is wasting money literally for anything]
One thing that would be super useful is a per-transaction approval prompt sent back to the human owner through chat, so I can quickly allow or deny a purchase without digging through a dashboard.
Honestly kind of wild seeing my agent check out on its own without me babysitting the checkout flow. The spend limits give me peace of mind though, would be a nightmare otherwise.
finally spun up a quick test agent and pointed it at a dummy checkout. it actually paused for confirmation before charging, which i was not expecting from a card pitched at bots.