YT YouTube2Text

YouTube2Text → Transcripts → How AI agents will change payments, commerce, and fraud - Amir Sarhangi (CEO, Skyfire)

How AI agents will change payments, commerce, and fraud - Amir Sarhangi (CEO, Skyfire) — Transcript

by Wrap Up · 5,590 words · 797 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00The projections are by 2030 a trillion
0:04dollars worth of commerce will be driven
0:07by AI agents. First step is just to even
0:09understand how often agents are coming
0:12to their website. Right? Today one of
0:14the challenges merchant uh merchants
0:16have is that they don't actually
0:18understand what that is. If you look at
0:20the statistics the amount of traffic on
0:22the internet has 10xed in the last
0:24couple of years, right? And over 50% of
0:27that is essentially bot traffic. Those
0:30types of actions will create a lot of
0:32problems when it comes to fraud and when
0:34it comes to security.
0:36>> Okay, guys. Machine economy, well, it's
0:38complex, right? I I totally understand.
0:41Okay, let me simplify it. AI economy or
0:44AI is making purchase on behalf of us.
0:47It is still um quite complex. I totally
0:49get it. That's why today I'm joined by
0:52Amir Sarangi. um CEO of Skyfire and I
0:57believe this is the right person to
0:58explain us what is um machine or AI
1:02economies, how AI is going to change,
1:05how we do purchases, how we do
1:07e-commerce and how we can safeguard
1:10ourselves when AI agents start making
1:12the payments on behalf of behalf of us.
1:15Amir, welcome to the show.
1:17>> Sam, thanks for having me. Great to be
1:19here. Right. Amir, just the first
1:22question straight. What is know your
1:24agent? If you could explain in a couple
1:26of words.
1:27>> Yep. Know your agent uh is actually a
1:30term that we came up with a few years
1:31ago when we first started the business.
1:34And really is synonymous with uh when
1:37you think of KYC, know your customer. So
1:40know your customer is about actually
1:42doing verified check and ID check on on
1:46a human being when you open a bank
1:48account. For example, uh similarly uh
1:51for KYA which is know your agent is
1:54understanding who really this agent is
1:56representing. So it's a combination of
1:59verified information on the human that's
2:02behind the agent,
2:03>> the agent developer itself. That agent
2:06developer could be a company like uh
2:08chatbt or anthropic uh or it could be an
2:12individual developer um with like
2:14OpenClaw or something like that. Uh so
2:16it's both of those two information
2:19combined in a token we call it KYA
2:21token. Uh and that know your agent also
2:24has information about the intent of the
2:26user. So for example if Sam is trying to
2:29purchase uh tickets using an AI agent uh
2:33has what is Sam trying to do and then
2:36has Sam given permission to that agent
2:39to do that on on his behalf.
2:41>> Right. And I mean look um I want to
2:45understand uh when a company develops an
2:48AI agent I understand like for example
2:50open AAI or some some other big
2:52companies but when an individual
2:54developer let's say I sit at home I
2:57develop my own agent and it starts
2:59making um a purchase on my behalf right
3:02at Skyfire how do you guys understand
3:06um that this is a legitimate bot you can
3:09trust it you can trust the information
3:11sending you guys what I'm trying to say
3:13that I'm trying to understand if there
3:15are certain rules I mean in the industry
3:17or you guys created perhaps that says
3:20yes you can trust this agent you can
3:22trust this AI or no you cannot try trust
3:25that how do you do that at Skyfire
3:27>> yeah great question Sam so that is
3:29probably one of the biggest uh problems
3:32that exists today uh is that it's very
3:35difficult to understand who these agents
3:37are if you look at the statistics um the
3:41amount of traffic on the internet has
3:43you know 10xed in the last couple of
3:46years right and a lot of that is over
3:4950% of that is essentially uh bot
3:52traffic
3:53>> okay
3:53>> now some of this bot traffic is
3:55legitimate some of it uh is actually
3:57potential revenue uh for a lot of the
4:00websites that these agents are trying to
4:02go to but a lot of it is also fraudulent
4:05as as you mentioned like a lot of times
4:07like you need to know who these bots are
4:09so what We uh you know what really we
4:13have been embracing is with KYA is this
4:16concept that uh bot developers need to
4:19go through a KYB process so that okay
4:22you actually know who this is. So in in
4:24the case of an individual or an
4:27enterprise uh they would go through a
4:29KYB which is know your business and
4:31that's proving that uh this is really
4:34who you say you are and then
4:36additionally uh when it comes to the end
4:39user having verified information on who
4:42that end user is becomes pretty critical
4:45because if you're trying to deal with a
4:46merchant or a business that business
4:48needs to know who you are uh and needs
4:51to be able to in order to be able to
4:53provide you the service that you're
4:55asking for.
4:56>> Yeah, true. And um before going deeper
5:00into this topic, could you briefly
5:01explain what you what you guys are doing
5:03at Skyfire? What pro problems you are
5:05solving?
5:06>> Yeah, so uh at Skyfire really there are
5:10three areas that we're focused on. One
5:12is uh agent identity. Uh secondly is
5:16around access. So how do agents access u
5:20>> basically human websites, right? like
5:22just be able to go to directly to to a
5:25website and be able to interact.
5:27>> And then thirdly, ability for to be able
5:29to uh check out. So for example, if uh
5:33an agent goes to a website, discovers a
5:36product it likes, how can it actually
5:38purchase it like a human being does,
5:40>> right?
5:41>> The challenge today is that most of
5:43these agents get blocked uh at different
5:46layers of web security. So for example
5:49um most merchants have uh blocking
5:52technology like Cloudflare, AI uh today
5:55in fact we announced with Akami and
5:56Experian.
5:57>> Um and then additionally you have fraud
5:59checks uh this is folks like forer who
6:02are at fraud
6:05>> basically inspecting to make sure this
6:06is a real human being. And then you know
6:09one of the other big things is that when
6:12you're purchasing uh a lot of times you
6:14want to be able to use your loyalty. So,
6:16you know, let's say you go to Best Buy,
6:18you want to be able to make sure that uh
6:20purchase that you're doing also counts
6:23for your loyalty. So, ability to log in
6:25becomes another big aspect and that's
6:28typically protected by folks like Octa.
6:31>> All right. And um where do you sit in
6:33this flow? I'm trying to understand. Do
6:35you directly work with merchants or you
6:37work with payment service providers?
6:39Where do you sit? And um should
6:42merchants go um make the integrations?
6:45How does it work? if you could explain
6:46as well.
6:47>> Yeah, great question. So, uh what we do
6:50is we work directly with the demand
6:52side. So, this is like an agent
6:53developer. An agent developer uses our
6:56technology in this case uh what we call
6:58a KYA token. They create this KYA token.
7:01That token has the verified information
7:04on the human and the and the agent
7:07itself. Uh and then that's used in the
7:10header. when an agent makes a request to
7:12a website that's included as part of the
7:14header.
7:15>> Uh we have gone out and done uh
7:18basically partnerships with a host of
7:20web security companies. Uh again this is
7:23people like ai um data dome right
7:26>> cloudflare and they uh these web
7:30security companies accept this uh token
7:32and then provide access to the websites.
7:35The
7:35>> beauty of this model is that it actually
7:38scales across the whole internet without
7:41all of the different merchants and
7:42websites having to support anything new.
7:46So there's no new requirements other
7:48than their web security uh company
7:51accepting this token.
7:52>> But but you just mentioned that for
7:54example yeah this is also not something
7:56new. Um for example let's say um agent
8:00developer needs to go and do KYB.
8:04Oh yeah agent developer. Now I now I get
8:06it. Yes. Yes. And um for agent developer
8:10to do the KYB this was my question I was
8:12thinking about let's say I'm an
8:13individual I don't have a business I'm
8:15just a developer who sits at my home and
8:18just develop it. How would I do a KYB?
8:20What would you require for me?
8:22>> That's a that's a good question Sam. So
8:23that basically turns into a KYC. So in
8:26the case where there is no there is no
8:29entity uh it becomes essentially a KYC
8:33um and that becomes an individual
8:35developer who basically is accessing
8:38something on it on on on his or her
8:41behalf right so yes but there's still a
8:44verification process and
8:46>> you still have to go through yeah
8:48>> exactly
8:48>> yeah and if you talk about the machine
8:51economy where do you see at the moment
8:55AI agents are being used the most. Is it
8:57like e-commerce or subscription economy?
9:00Where is this happening the most at the
9:01moment?
9:02>> Yeah. So look, I mean purchasing with
9:05bots has something that's been happening
9:07for many years already. It's not new.
9:10>> So for example, if you if you look at,
9:12you know, the reason why when you want
9:15to go to a really popular concert and
9:17within seconds,
9:19>> uh, most of it is sold out even though
9:20you're refreshing your web page. It's
9:23not just because there's a lot of fans
9:25trying to buy. It's also there's bots
9:27that are coming in and trying to
9:29purchase on behalf of resellers. Same
9:32thing is true with Nike drops and many
9:34other places in e-commerce. So,
9:37>> right,
9:37>> bots have been purchasing, you know, I
9:40wouldn't say legally, but um you know,
9:42in in a gray zone manner. So that's
9:45something that's already been happening
9:46it and you have companies like now that
9:49are much more you know front and center
9:52about it like rye one of our partners
9:54card AI there's a bunch of uh players
9:57that are already now purchasing uh
10:00either on behalf of businesses or uh
10:02consumers
10:03>> that's already happening on the machine
10:05economy side of things when it comes to
10:07micro payments I'd say it's still early
10:09days but we're starting to see um
10:12basically ability for agents on on their
10:15own to be able to pay for content and
10:18also be able to pay for APIs or access
10:21to uh whatever it is that they're trying
10:23to get to. But I'd say, you know, from
10:26our perspective, we see that still at
10:27early days. Part of the challenge there
10:30is that a lot of the companies on the
10:33supply side of this equation continue to
10:35want to be able to sell SAS, right? It's
10:37it's much better to sell a SAS package
10:40than a ondemand per uh request basis. Uh
10:44but that is going to change. Uh and the
10:47reason that will change is because you
10:49have many companies who will um not want
10:53to pay the big SAS dollars especially if
10:55they're not using the full package.
10:57They'll go and try to you know use
10:59upstarts or even create their own you
11:02know stack. Um so for those reasons
11:05those pressures are going to force um
11:08those SAS bundles to uh be unbundled and
11:12be offered on a more uh alakart basis.
11:15>> You just you just mentioned uh microp
11:17payments and I often hear that it is one
11:19of the key parts of machine economy. Why
11:22are microp payments so important in this
11:24machine economy? What role they play?
11:26Yeah, I mean it basically if you want to
11:29now let's say uh let's say you use
11:31HubSpot or you or okay your agent wants
11:34to have access to some particular
11:36content uh for learning or for it just
11:40needs access to it and wants to be able
11:42to pay for that content on a per article
11:44basis or wants to have access to an API
11:48uh for for whatever it is it needs.
11:51Yeah,
11:52>> that API call is a single call and a
11:55single access and it may come back a
11:56hundred more times and use it again. But
11:59each time it's going to pay a cent or a
12:02fraction of a cent for that access. And
12:04that's where that microp payment becomes
12:06important because uh in in this
12:09scenario, the agent is not signing up to
12:11some big plan. It's saying I'm going to
12:13pay you every time I use your product.
12:15Which means that the the provider has to
12:18be able to sell its product on a uh
12:21essentially on a single hit basis.
12:24>> Right. Right. And as you know um current
12:26rails are not best for micro payments.
12:29Um those are very expensive. Card rails
12:32especially they are very expensive.
12:33That's why everyone is trying to avoid
12:35that. Do you think that's why um stable
12:38coins would become like a default
12:40currency for machine economy or do you
12:43think um the current rails adopting and
12:46they will still exist in one or another
12:48way?
12:50>> Uh I think look I think you're going to
12:52see both but I this is where the sweet
12:55spot for uh stable coins is because it's
12:58not just the micro payment aspect of it.
13:00It's the fact that you're actually
13:02transferring value at the time of the
13:05service. So imagine a scenario where you
13:08hire another agent that then hires
13:10another agent to do something
13:12>> in that chain. You know, you have an
13:15agent that's getting a commission or
13:16getting part of a transaction. Now
13:18imagine there's fraud or you know this
13:22can't be done in credit because like
13:23trying to roll that back or
13:26>> figure out who caused what fraud becomes
13:28a big hassle, right? Whereas here with
13:31stable coins once that money is
13:34transferred like you know that that next
13:37agent owns that you know um that
13:42currency which is that becomes really
13:44critical that u that value exchange uh
13:48at the time of service is is pretty
13:49critical when it comes to kind of this
13:51new agentic world that we're going to
13:53see.
13:54>> I see. And do you guys do transactions
13:56actually you or you do not touch money
13:59at all? You are only working on as you
14:01said as a know your agent and basically
14:04assisting the transaction.
14:06>> Yeah, we provide a wallet. In fact, we
14:08started with stable coins.
14:09>> Uh now we have with our partnerships
14:12with Visa, Mastercard, we've added uh
14:15both uh Visa tokenized cards and
14:18Mastercard and we'll we'll be also uh
14:21adding others. Um so our wallet
14:25basically will provide the agent all the
14:27different mechanisms uh for payment
14:30essentially.
14:31>> Okay. Okay. And uh why should I use your
14:34world for example? Why cannot I go um
14:38use only your KYA let's say know your
14:40agent and continue using Stripe or any
14:43other service provider that I'm using at
14:45the moment.
14:46>> Yeah. I mean and you certainly can. The
14:48big advantage of using us uh other than
14:50being obviously a one-stop shop for you
14:52is the fact that when an agent is trying
14:56to go to a a website or to uh an API,
15:01one of the important things that that
15:03end destination wants to know is that
15:06there's real intent for purchasing.
15:08meaning okay we can say that like Sam
15:10the Sam agent
15:13>> has a credit card on file and okay is a
15:16real buyer right and then at the time
15:20that you do try to make a purchase um
15:23understanding not only that this is not
15:25a fraudulent card that you're using
15:27that's something that we we're working
15:29with Experian to unlock is this kind of
15:32real time risk
15:33>> uh aspect um and then secondly is is the
15:37fact that uh we capture the intent of
15:41the user in that process meaning that
15:42you've given uh authorization
15:46uh to the agent to make that purchase on
15:48your behalf. All of that together
15:50becomes really critical for uh merchants
15:53who are trying to provide you know SAM
15:57AI for example AI agent uh a service or
16:01a product.
16:02Does it mean that well basically if I
16:04understand correctly if I take the
16:06bundle from you meaning like your wallet
16:08and other services it will be more
16:11secure than combining it with let's say
16:14other providers that I'm already using.
16:16Do I understand it correctly? it not
16:18only more secure but also you get uh
16:21ability to go to uh probably more uh
16:25destinations and websites because now we
16:28have a complete uh understanding of who
16:30this agent and the user behind it is and
16:34the intent of that user. So it's really
16:36becomes a combination of access and
16:38security in that scenario. And what are
16:42some of the um industries that are
16:44adopting AI agency in the first place?
16:47And here's the interesting question. How
16:50do you guys make money out of this?
16:53>> Uh well, we we make money in a few
16:55different areas. Um but really initially
16:58really it's about adoption. So um where
17:02we make money is KYA tokens being
17:04created. Um and that's something that
17:07you know we work uh directly with agent
17:10developers uh with uh but also in the
17:13future you know we'll hopefully also be
17:16able to uh provide more risk signals to
17:20the merchants and that will be another
17:21area. Um in terms of you know we're
17:24seeing quite a few different areas in
17:27terms of um adoptions uh we are working
17:31uh with uh not only kind of AI first
17:35vertical um developers who are
17:38essentially focused on things like
17:40whether it's fashion shopping uh travel
17:44you're going to see a lot more
17:45personalized AI assistant type
17:48experiences um that are going to be um
17:51essentially providing a really high-end
17:54experience to uh to users. And then the
17:58other area that we're uh getting some
18:00great traction around is financial
18:02institutions. Um this is card issuers.
18:05This is um folks who are essentially uh
18:09providing uh you today a wallet. Uh and
18:13and that wallet could be either through
18:16uh your your card or your bank. Uh so
18:20that's another area that we're starting
18:21to see uh quite a bit of interest.
18:23>> Amir, what do you think? Which
18:25industries are adopting AI agents the
18:28first? Where do you see the real
18:29adoption curve basically going up?
18:32>> Yeah. So in the consumer space, uh I
18:35mean there's consumer and B2B. On the
18:36consumer side, what we're seeing is that
18:39you have uh AI first companies that are
18:42being created now. Those companies are
18:44starting
18:45>> really with a completely new user
18:47experience. this is more of a, you know,
18:50kind of a chat commerce experience.
18:52That's uh, you know, we're starting to
18:54see that in fashion. We're seeing that
18:56in travel um, and you know, many kind of
19:00different categories in the consumer
19:01space. You're also starting to see
19:05existing companies uh, who previously
19:08are like a marketplace. uh this is where
19:10you go to discover new things like think
19:12Pinterest or others like that who are
19:15also uh wanting to take the friction out
19:18of the buying experience. So anywhere
19:20where you have a referral link where
19:23that referral link takes you to another
19:25site that can be now more of a buy for
19:27me button. It's kind of like that
19:29instant purchasing experience where you
19:31now don't have to go to these additional
19:34uh steps and websites to to purchase. So
19:36that's more on the consumer side. Um,
19:38we're also starting to see financial um,
19:41>> companies, you know, whether it's uh
19:43your card network, anyone anywhere where
19:46you have a wallet like Think Chase
19:48Preferred or uh, you know, Capital One,
19:52those types of experiences will evolve
19:54in the future and you know, you will now
19:57just be able to say, "Hey, this is what
19:59I'm looking for." It will find you the
20:01best price, apply the coupons, and be
20:04able to make that a much more of a
20:06concierge. uh kind of a VIP experience
20:09for you. Do you do do you think
20:13I mean we are in agent or AI economy or
20:17it's very very early stage why I'm
20:18asking this question is that you must
20:20have read that open AAI and stripe they
20:23launched a partnership so that they
20:25enable purchasing directly inside the
20:28chat but but then they scaled back this
20:32partnership um there was a news so why
20:35do you think that happened in your
20:37opinion
20:39>> um there There there are many different
20:41reasons um for that you know I obviously
20:44I I'm not inside OpenAI I cannot give
20:47you
20:47>> sure
20:48>> anything but from a outsers's
20:50perspective you know I think um the
20:52experience was not there yet you know
20:55from a from a consumer side uh and
20:57really you know if you think about it
20:59from a consumer perspective when you
21:02want to use a AI agent part of the
21:05reason you do this is because you wanted
21:07you expect that agent to do a better job
21:09than you it, right? Like if it's just to
21:11go to Walmart to buy something, you'll
21:13just go directly to Walmart and buy it.
21:15Like you don't need an agent to do that
21:17for you, right?
21:18>> Where this really powerful is that, hey,
21:21you're looking uh for a new set of
21:23headsets.
21:25>> You figured out you want Bose and you
21:28just tell your agent, go find me the
21:29best price. If there are any discounts,
21:32offers, apply it and just buy it for me.
21:35that becomes more powerful when you know
21:37that that agent has done a much better
21:39job than you will
21:41>> you know scouring the internet to find
21:43that best deal and then just
21:45>> buying it regardless of whether you have
21:47a relationship with that merchant or not
21:50and I think that's where you know at the
21:52end of the day you know any of these
21:54experiences for them to work really well
21:58they need to be uh scaled across the
22:01internet and then more importantly there
22:03are other implementation aspects that
22:06have to be considered around fraud um
22:08and risk that you know um become really
22:13important in any implementation. So uh
22:16you know
22:18your question is a is a really good one
22:20uh but I think a big part of the reason
22:24um there was a pause in that particular
22:27instance is primarily because I think
22:29they rushed into it very quickly to to
22:31launch something. So you you think uh we
22:34are very early stage and there is no
22:36trust because again I have another
22:38example is Walmart. They also rolled out
22:41basically um this similar types of
22:44solutions but they rolled back and the
22:46news said that basically customers were
22:49not converting from chat to their
22:52website you know at all. It was like
22:54conversion rates were very very low. So
22:57they said like it we are rolling this
23:00back. Do you think this is like trust
23:01issue? uh people do not trust AI agents
23:04or there is something else.
23:06>> I think it it's trust definitely is a
23:09factor in terms of adoption of any new
23:11technology. Even if you look at the
23:13early days of uh you know commerce on
23:16the web, people would not trust using
23:20their credit card on a website. That was
23:23not too long ago. That that was just
23:25>> you know three decades ago when you know
23:28Amazon was trying to push people to use
23:30their card. Right. And
23:31>> right
23:32>> that evolution is one where you have to
23:35ease the user into that experience. Uh
23:39>> the way we see it is like three
23:41different steps. The first step is this
23:43experience where you see something, you
23:46like it and you say just buy it for me.
23:49Right? So that's a place where you've
23:51already discovered what you want and you
23:53make that purchase. So that that's the
23:55first way of easing the user into the
23:57experience. The second step is more of a
23:59commerce chat experience where you don't
24:02know what you want yet um and you're
24:04trying to figure out um what's the best
24:07option for you. So you start chatting um
24:10kind of like what you do search but here
24:12it's a much better experience because
24:14you can actually um do it in natural
24:16language
24:18be able to quickly find what you want um
24:21just like you do today in open AI. I
24:23mean the statistics are kind of
24:25astounding how much of these uh now you
24:29know uh referrals are coming from uh AI
24:32companies. The numbers are pretty
24:34astounding and the projections are by
24:382030 a trillion dollars worth of
24:41commerce will be driven by AI agents and
24:44and those numbers will happen very
24:46quickly. That's just three years from
24:47now. And the reason is that you're going
24:49to have this progression where you start
24:51from this kind of buy forme experience
24:53to agent kind of um chat experience and
24:58that still come into the loop. You're
25:00you're still approving everything. So
25:02when the agent goes and finds you what
25:04it wants, it'll ask Sam, "Hey Sam, uh do
25:07I have approval to buy this pair of Nike
25:08shoes for $110?" You approve it with
25:11your pass key and then goes and
25:13purchases for you. So, as you can see,
25:15it's just slowly building trust. And
25:17then kind of the third phase of this is
25:20more autonomous where you want to buy a
25:23pair, you know, you want to buy some uh
25:26concert tickets. You don't want to uh
25:28stay awake till midnight when they go
25:29live. You tell your agent, hey,
25:33if if you can find me this tickets under
25:35$200, just buy it. I approve it like
25:38beforehand uh with you. that becomes
25:42kind of the more autonomous experience
25:44and that's what we're starting to see
25:47and yes you are going to have
25:49implementations that are not going to be
25:51the best ideal user experience and they
25:54will need to be rethought but there's
25:56absolutely no question in my in my mind
25:59now that agent commerce is going to
26:01happen because we're starting to see it
26:03already working uh and more importantly
26:06it just gives you know in in some cases
26:10people want to shop. In a lot of cases,
26:12people don't want to shop and they much
26:14rather take that friction out and that's
26:16where uh agentic commerce will help uh
26:20free up time for people, make it much
26:23easier for them to do their daily tasks.
26:26>> I mean I understand for end consumer
26:28okay um but I I will not go that far to
26:32the future. You are saying to certain
26:34sense we are already in machine economy
26:36machines are already making purchases.
26:38Then what do you think merchants should
26:40understand about machine economy? How
26:43they should be prepared for that? What
26:46should they think about it and consider?
26:49First step is just to even understand
26:52how often agents are coming to their
26:54website right today. One of the
26:56challenges merchant uh merchants have is
26:58that they don't actually understand what
27:01that is. And and the reason I say this
27:03is that when you look at because
27:06obviously merchants are driven by data,
27:08right? Like data is king for them.
27:10>> And one of the things they care a lot
27:12about is understanding how often humans
27:14are coming to their website, what are
27:16they searching for,
27:17>> how often they leave a basket uh with,
27:20you know, goods in it. Um and now
27:24they're starting to see a big part of
27:26that pie chart of humans being gray.
27:29That gray is basically a combination of
27:32fraudulent agents and good agents that
27:34are coming to their website to try to
27:37understand their product or try to
27:39purchase something,
27:40>> right?
27:40>> They need to understand that that gray
27:42slice much better than they do today.
27:44And that's that's the first thing they
27:46need to be able to do is know who what
27:49type of agents are coming to them on
27:51behalf of what type of humans. And then
27:54being able to essentially unlock agents
27:57being able to purchase directly from the
27:59human website, no APIs. We're just
28:01talking about the regular old website
28:03because agents can do that now very
28:05effectively.
28:07>> And being able to unlock that, which
28:09then gives them much better
28:11understanding of uh what agents are
28:15driving revenue for them and being able
28:18to kind of foster that. and um and then
28:22being able to build essentially a a
28:23strategy around how they support agents
28:26going forward.
28:28>> What about financial institutions? We
28:31see you explained what should merchants
28:33do. What about financial institutions?
28:35Because they also need to be prepared,
28:37right?
28:38>> Yeah. So on the on the financial
28:39institution side, I think there's a
28:41couple of things. One is um you know you
28:44if you're a card issuer um if if today
28:48you are offering as a financial
28:50institution you are offering an
28:51experience to end us and consumers um
28:54the thing what that's really like this
28:58AI you know wave does for you is that it
29:01kind of evens the playing field it
29:03allows you to provide a much better
29:05experience to your users um that is very
29:08differentiated in the sense that now you
29:09can provide night. Uh, you know, let's
29:12say you're a card issuer, you have a
29:15wallet. Now, you can provide kind of
29:17this like really uh great concierge
29:20experience for your users to be able to
29:23use your card on the purchases that they
29:25want to do. So, it's, you know, think
29:28about it as kind of Amazon Prime where
29:30like once you're locked in and starting
29:32to use Amazon Prime for one thing,
29:35you'll start adding more and then it's
29:37just on automatic. It's the same story
29:39here. Once you start using an AI agent
29:42to do purchasing and you and that your
29:45card issuer basically that's the default
29:47card you're going to be using. Um then
29:50you are starting to really get your
29:53consumers to use your card for a lot of
29:56their purchases that are happening with
29:57a uh with agents. So that's one part of
30:00the stat strategy is really leaning into
30:04uh this new trend that will happen. Um
30:08and and the first step is just you know
30:10kind of that buy for me experience and
30:13then um from there more of a concierge
30:15experience. The second area where
30:18financial institutions are starting to
30:20need and and needing to evolve very
30:23quickly is that their customers want to
30:25be able to access their accounts using
30:28agents. So for example, okay,
30:30>> you know, being able to uh uh access
30:34your bank account or access your
30:36financials even in a readonly fashion
30:39using claude or chat or Gemini so that
30:43you can do analysis on your account for
30:46various right so right and similarly
30:49enterprises who are automating they want
30:52access to their own accounts using
30:54agents. So this becomes another big area
30:56for financial institutions. This is not
30:58so much agentic commerce but it's more
31:00like agentic access. So how does it
31:03>> right
31:04>> how do I access my own account using
31:07agents essentially that's the other area
31:08for financial institutions.
31:10>> I mean this is my last question. What do
31:13you think the biggest risk in the near
31:15term in machine economy? What should
31:18both merchants and financial
31:20institutions
31:21be ready?
31:23Yeah, look I you know I think in some
31:26ways if done right we can actually
31:29reduce fraud believe it or not u in
31:32agent with agent commerce because you
31:35have a much more secure way of
31:38onboarding the user and their card and
31:42then the essentially the way um
31:47everything moves from there is can be
31:50done in a very secure way I I think the
31:53challenge here and the risk is going to
31:55be if um we don't do that and we you
31:59know we have things like you know people
32:01using their you know raw pan their
32:05credit card number directly with agents
32:08or
32:09>> users trying to use an agent with their
32:11own username password
32:13um trying to access their bank and doing
32:16like two off with their own phone that
32:20those types of actions will um will
32:23create a lot of problems when it comes
32:25to fraud and when it comes to security.
32:28>> So it it it's more about actually
32:32you know the industry moving quickly
32:33enough to be able to provide the right
32:37tools uh for both agent developers uh
32:41and also enterprises to be able uh to do
32:45this much more securely.
32:47>> Right air really enjoyed our
32:49conversation. Thank you for coming on
32:50the show and sharing this insights. I
32:52hope it was useful for my audience as
32:54well. For everyone listening this
32:55episode, don't forget to share it with
32:57your friends and everyone interested in
32:59this. And don't forget to subscribe.
33:01Thank you very much.
33:03>> Thanks, Sam.

About this transcript

This page contains the full transcript of How AI agents will change payments, commerce, and fraud - Amir Sarhangi (CEO, Skyfire) by Wrap Up, generated from the public captions YouTube serves with the video. The transcript has 5,590 words across 797 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.