Darren Zuckerman

one of the more ironic things about high-risk payments is that merchants paying the highest processing fees usually get no tools to prevent fraud. Stripe merchants get Radar. high-risk merchants should expect the same. so we built Sonar at @frame_payments. docs.framepayments.com/conce…

my admitted skepticism of agentic commerce made it into a @FINTECHTVglobal interview today. there's a lot to figure out. refunds, exchanges, or disputes happen, who owns these problems? @scottew's recent experience is a perfect example of how messy all this can get: nitter.gq/scottew/status/2117013…

gas stations realized the real money wasn't in gas. it was in everything sold around it. payments is having its convenience store moment. if you're just moving money and charging a fee, you're in trouble. wrote about why payments companies are dying.

we've been shipping a lot of improvements to Frame lately, but one of the biggest is coming this week. one piece of feedback we've heard was that our dashboard was confusing. payments and compliance were grouped together in ways that made sense to us internally, but not always to the platforms using it. fair criticism... so we fixed it. the new dashboard gives each product a clearer home and makes it easier for us to introduce new features across financial services and compliance. we're rolling out launching v2 of our Transfers API. it supports bidirectional money movement across cards, bank accounts, and wallets. oh, and accounts on Frame can now hold balances too 👀 a lot more coming.

really honest question. if the agent still needs approval at checkout, what friction have we actually eliminated? with that said, cool to see AMEX experimenting with authenticating purchase intent.

been fascinated with OTAs and how they pay hotels whenever you make a booking. one of the more common flows goes like this: OTA (like Expedia, Booking, etc) charges your card, issues a virtual card and emails it to hotel, hotel charges that card around check-in. two card transactions, each with processing costs + plus virtual card issuance. hotels want payment assurance at check-in without taking on the risk of ACH returns. so we’re using card networks to move money between two parties that already have a commercial relationship. feels like there could be a better way?

I got into payments basically by accident. I was working with an OnlyFans competitor doing hundreds of millions revenue, paying our processor ~15% per transaction. the limitations of our processor forced me to learn payments from the ground up. CCBill had no API. no fraud monitoring. no customer profiles. no stored payment methods. we ended up building our own fraud tooling because we had no choice. I didn't know enough about payments back then to know if we could use a gateway like Authorize or NMI to improve the experience. we sold video content, so we learned fast that videos not playing meant chargebacks. video infrastructure suddenly became a payments problem. after paying 15%, this was our checkout experience. no Apple Pay. no address autocomplete. most users dropped off here lol CCBill alone processes over billions per year. but there’s a mountain of volume trapped with legacy processors in high-compliance industries. we plan to eat our way through that market on the way to Stripe.

we started building frameOS in January. ~90 days later, our first accounts were live. six months ago we had 16 accounts. now we're over 60,000. processing volume driven by these accounts is up 895% in the last 60 days alone. the idea behind frameOS is simple. compliance and financial services have operated in separate silos, despite being dependent on each other. we built them into the same system. an account can represent a user, business, or platform. each has capabilities to move money across cards, bank accounts, and wallets, which unlock as verifications like KYC, KYB, adverse media, and geo-compliance pass. we're powering use cases across prediction markets, creator platforms, telemedicine, vertical SaaS, ISVs, and more. compliance shouldn't sit alongside financial infrastructure. it should be part of the infrastructure itself.

mood heading into the weekend after the week we just had at @frame_payments . a lot coming over the next couple weeks. new customers going live, new features shipping, and some of the coolest stuff we’ve built yet.

as we go deeper working with these platforms, one thing I keep coming back to is how much signal we see. we see more than most processors because we also handle compliance for a lot of these platforms. we see payouts patterns, spend behavior, location, and the identity data. when you put this together, it’s a fuller picture of a user. it can flag when someone might need limits imposed before it becomes a real problem. it's a realllllly fine line though. protecting users from fraud is our job. protecting people from themselves isn't. we're not deciding how someone spends their paycheck, but we can make sure platforms have the information to make that call themselves. the signal only matters if it turns into something they can act on.

the struggle described in this piece makes sense. as a founder, your instinct is to remove friction and grow. I think there's a happy medium between being overly sensitive to risk and being reckless about it. prediction markets are a unique fraud surface because a win pays out in cash. we process payments for one at Frame and see the same patterns. stolen cards, card testing, velocity attacks, etc. we think the approach should be to push verification upstream. we combine device signals, KYC, $0 3DS auth, card/address name matching, and more into onboarding to tie the user to the payment method before they transact. this is what a good UX looks like…verification happens before the first transaction so you have fewer actual fraud losses (friendly fraud is another story). wsj.com/business/polymarkets…

we’re sponsoring a builder house in Austin. mostly to make sure nobody says “we’ll just use Stripe.” sign up to attend and we’ll throw ya $100k in processing credits for whatever you’re building.