WEBVTT
1
00:00:10.180 --> 00:00:11.013
How is everybody doing?
2
00:00:12.320 --> 00:00:13.153
All right.
3
00:00:13.240 --> 00:00:16.700
Really? That's not good. Let's ask that question again.
4
00:00:16.820 --> 00:00:19.820
How is everybody doing? Yeah.
5
00:00:20.920 --> 00:00:24.020
It's a thousand-people room. We need to make it look like thousand-people rooms,
6
00:00:24.120 --> 00:00:28.680
so let's get the energy out. I'm Kalyani and this is Danny.
7
00:00:28.860 --> 00:00:33.580
We'll get to us a little bit in a minute, but it's all about you guys.
8
00:00:34.220 --> 00:00:39.140
How many of you have used an AI agent to do anything in the last
9
00:00:39.220 --> 00:00:40.500
one week? Please raise your hands.
10
00:00:42.320 --> 00:00:44.840
And those of you who haven't raised your hands,
11
00:00:44.840 --> 00:00:47.860
I know you have used it.
So how about,
12
00:00:49.100 --> 00:00:54.060
have you thought of anything in the last one week or last one month or one
13
00:00:54.720 --> 00:00:56.820
year-is there a task that you said,
14
00:00:56.820 --> 00:01:01.260
"I wish something somebody can automate it end to end
15
00:01:01.580 --> 00:01:06.540
that I don't have to do this anymore?" Please raise your hands if you have
16
00:01:06.580 --> 00:01:10.640
at least one or two tasks that you want end to end. Okay. That's fair,
17
00:01:10.720 --> 00:01:13.340
good enough. Now let me ask you one more question.
18
00:01:13.520 --> 00:01:16.440
I promise this is the last question. Maybe I'm lying. But
19
00:01:18.360 --> 00:01:21.900
what if that agent needed to pay to complete your task?
20
00:01:22.160 --> 00:01:25.760
Would you let that agent do that? Okay.
21
00:01:26.040 --> 00:01:26.873
Please say yes.
22
00:01:27.020 --> 00:01:29.680
Yeah. No,
23
00:01:29.900 --> 00:01:33.040
but I think that's what we are here to talk about, right?
24
00:01:33.380 --> 00:01:37.960
How do you go from using the agent to do productivity tasks to
25
00:01:38.180 --> 00:01:40.560
using and pushing the boundaries on it?
26
00:01:40.660 --> 00:01:42.260
We have all the tools and everything else.
27
00:01:42.820 --> 00:01:45.060
We've been hearing about it in the last couple of days,
28
00:01:45.480 --> 00:01:50.160
but how do we go from using the agent to write your emails to actually
29
00:01:50.340 --> 00:01:53.740
buying something for you? And last question:
30
00:01:54.700 --> 00:01:59.380
how many of you work for a company or know of a company which will accept
31
00:01:59.880 --> 00:02:03.680
a buyer agent? Is there anybody who's working at-oh, wow.
32
00:02:04.080 --> 00:02:07.120
You guys are full friend. Awesome. So
33
00:02:09.180 --> 00:02:13.280
we are here to talk about that.
And just to give where we are in the
34
00:02:14.460 --> 00:02:15.580
agentic ecommerce space,
35
00:02:16.720 --> 00:02:21.100
payment industry has been building for agents because technology has advanced to
36
00:02:21.120 --> 00:02:25.500
that stage where agents can have wallets. 12 months ago,
37
00:02:26.060 --> 00:02:30.380
this space had nothing. They were just starting to talk about standards.
38
00:02:30.440 --> 00:02:31.740
Now we have more than 10 standards.
39
00:02:31.800 --> 00:02:36.520
We have agents who can have wallets and agents can now complete
40
00:02:36.540 --> 00:02:38.600
purchases without a human click.
41
00:02:39.040 --> 00:02:42.520
What I mean by that is not that human is not involved or we are not
42
00:02:42.600 --> 00:02:46.300
involved-humans sounds very distant-we are not involved.
43
00:02:46.540 --> 00:02:50.760
It's just that we tell the agent the policy, we set the governing rules,
44
00:02:50.940 --> 00:02:55.060
and then we don't have to put in our credit card information and everything else
45
00:02:55.160 --> 00:02:58.780
into a checkout form. The agent can take care of that for ourselves.
46
00:02:59.140 --> 00:03:01.280
And that's where we are now. 12 months ago,
47
00:03:01.640 --> 00:03:03.800
every transaction had to go through a checkout flow.
48
00:03:04.620 --> 00:03:09.080
And this is, again, 12 months ago,
49
00:03:09.140 --> 00:03:12.020
there were a lot of prototypes, but now we have actually production.
50
00:03:13.220 --> 00:03:17.680
I know there's 100 million agent transactions reported according to the sources.
51
00:03:17.800 --> 00:03:21.800
And yesterday, if you heard the keynote,
52
00:03:22.300 --> 00:03:27.080
there were so many features that Stripe launched that help agents
53
00:03:27.400 --> 00:03:28.980
have a wallet, have a rules,
54
00:03:29.200 --> 00:03:32.300
and you as a businesses have bot protection as well.
55
00:03:32.300 --> 00:03:34.960
So to be able to recognize when there's a bot abuse.
56
00:03:35.060 --> 00:03:36.760
So there's a lot of things that are happening today.
57
00:03:37.180 --> 00:03:40.460
So we're going to talk about it.
So I'm Kalyani Kopisetti.
58
00:03:40.840 --> 00:03:43.200
I'm a partner solutions Architect at AWS.
59
00:03:43.300 --> 00:03:47.820
I've been in the financial services industry for 25-plus years and my
60
00:03:48.640 --> 00:03:51.560
recent focus has been payments and
61
00:03:54.260 --> 00:03:57.340
agentic ecommerce and I'm here with Danny. So, Danny,
62
00:03:57.780 --> 00:03:59.000
take it away and get us started.
63
00:04:00.060 --> 00:04:01.540
Everybody, my name is Danny Smith.
64
00:04:01.940 --> 00:04:04.600
I am the agentic commerce solution architect lead at Stripe.
65
00:04:05.440 --> 00:04:07.240
I apologize in advance for my voice.
66
00:04:07.640 --> 00:04:10.900
It's just due to all the many exciting conversations I've had all week with
67
00:04:11.200 --> 00:04:15.560
users, partners, people in training. So I'm going to push through,
68
00:04:15.780 --> 00:04:19.440
but very happy to get us rolling.
69
00:04:19.680 --> 00:04:22.980
So there is an ecommerce paradox.
70
00:04:23.360 --> 00:04:24.860
You probably heard this a few times this week.
71
00:04:25.580 --> 00:04:28.100
Most websites are built and optimized for humans.
72
00:04:29.480 --> 00:04:32.500
And if you look at the story of this bot right here, actually,
73
00:04:33.240 --> 00:04:34.440
I feel like this bot is,
74
00:04:35.060 --> 00:04:36.980
he's hit a brick wall like my voice has hit a brick wall.
75
00:04:37.820 --> 00:04:42.220
What is that brick wall? There's a file called robots.txt.
76
00:04:42.560 --> 00:04:47.080
Show of hands. And I couldn't let you use all the questions. So, show of hands,
77
00:04:47.500 --> 00:04:52.440
who's heard of robots.txt? Looks like about half. Okay.
78
00:04:54.000 --> 00:04:55.220
So basically,
79
00:04:56.320 --> 00:05:01.040
there's this paradigm in agentic commerce where your high-intent,
80
00:05:01.580 --> 00:05:06.340
agentic buyer and agent may actually be blocked at the front door or hit a brick
81
00:05:06.420 --> 00:05:10.400
wall due to how you have your tech stack configured,
82
00:05:10.500 --> 00:05:13.560
how you have your website configured.
So this is the paradox.
83
00:05:14.220 --> 00:05:16.280
We don't want to turn away these high-intent buyers,
84
00:05:16.620 --> 00:05:19.360
but we also want to make a safe and secure experience.
85
00:05:19.780 --> 00:05:24.500
So this bot and this example is going to get turned away and it's going to go on
86
00:05:25.180 --> 00:05:28.000
to the next website. It doesn't call customer support.
87
00:05:28.080 --> 00:05:31.640
It doesn't fill out forms. It just moves on to the next.
88
00:05:32.200 --> 00:05:36.300
So this is a high intent buyer that we actually care about in agentic commerce
89
00:05:36.800 --> 00:05:40.860
and we want to basically facilitate that buyer as we move forward.
So what we're
90
00:05:40.900 --> 00:05:44.760
going to do is show you a prototype. Okay? This is not a production.
91
00:05:45.400 --> 00:05:46.480
This is showing the flow.
92
00:05:47.120 --> 00:05:49.580
This is an accurate flow of where we're moving in the industry.
93
00:05:50.220 --> 00:05:53.000
So I have a buyer's agent. I'm giving it a task.
94
00:05:53.440 --> 00:05:58.360
And this task is basically "Generate an image for a sneaker marketing launch
95
00:05:58.400 --> 00:06:02.520
campaign," and I'm going to give it the astronomical budget of $2
96
00:06:04.220 --> 00:06:09.120
to accomplish this task. Now, that's a lot of money in microtransactions.
97
00:06:09.120 --> 00:06:10.820
This is living large, two whole dollars.
98
00:06:10.920 --> 00:06:14.920
We're going to go forward and we're going to contact three websites and we're
99
00:06:14.960 --> 00:06:17.440
going to move forward and basically say,
100
00:06:17.900 --> 00:06:22.440
"Go build this image for me as part of this campaign." So seller one
101
00:06:23.160 --> 00:06:25.020
rejected. Actually,
102
00:06:25.100 --> 00:06:28.280
the seller had no visibility to this transaction even happening.
103
00:06:28.340 --> 00:06:30.800
It doesn't even know that a buyer was at the front door.
104
00:06:31.160 --> 00:06:33.480
It was immediately rejected. And we're going to talk about why in a second.
105
00:06:34.080 --> 00:06:38.040
Seller two, a little bit better. It read the manifest.
106
00:06:38.100 --> 00:06:39.520
It got to the product catalog.
107
00:06:39.880 --> 00:06:43.520
It got to a point where it was trying to complete an actual transaction.
108
00:06:44.200 --> 00:06:46.460
The payment rails were not in place.
109
00:06:47.560 --> 00:06:51.320
And seller two came back to this buyer agent and said, "Hey,
110
00:06:51.860 --> 00:06:55.360
sign up for an account." When an agent tells another agent to sign up for an
111
00:06:55.380 --> 00:06:59.800
account, it's not going to happen. So it's going to move on.
Seller three,
112
00:07:00.760 --> 00:07:05.320
the entire experience was automated by a seller's agent.
113
00:07:05.680 --> 00:07:09.880
So step one, it found the website. It got through the brick wall.
114
00:07:10.700 --> 00:07:12.840
Step two, it found a manifest,
115
00:07:12.900 --> 00:07:15.200
and we're going to walk through exactly how this happens.
116
00:07:15.820 --> 00:07:19.940
A manifest is telling this agent, "Here's how you transact. Step three,
117
00:07:20.820 --> 00:07:25.200
the payment rails were in place." So this was actually a good experience and
118
00:07:25.320 --> 00:07:26.160
end-to-end experience.
119
00:07:26.600 --> 00:07:30.700
The transaction actually happened and we got it done in three seconds
120
00:07:31.720 --> 00:07:36.560
for 12¢. So way under budget, everyone's happy. So again,
121
00:07:36.780 --> 00:07:40.620
it's one product that we're trying to find, three different sellers.
122
00:07:41.380 --> 00:07:46.120
Two sellers had no visibility to this even being an enabled buyer.
123
00:07:46.660 --> 00:07:50.960
The third seller was able to get that visibility, actually transact,
124
00:07:51.920 --> 00:07:54.640
paying customer, everyone's happy. So we're going to walk through,
125
00:07:55.040 --> 00:07:55.873
kind of in detail,
126
00:07:56.240 --> 00:08:00.960
how we can get to that third seller modality to make sure that agent-to-agent
127
00:08:01.000 --> 00:08:04.840
commerce becomes a reality as we move forward.
So we want to focus first on the
128
00:08:04.880 --> 00:08:09.520
buyer's agent. So Kalyani, give us more background on the buyer's agent.
129
00:08:09.880 --> 00:08:10.520
Yeah.
130
00:08:10.520 --> 00:08:14.560
So as most of you didn't raise your hands because you didn't want to give a
131
00:08:14.680 --> 00:08:17.400
buyer agent your credit card, assure you,
132
00:08:17.760 --> 00:08:20.840
there's agent wallets that you can give them with $2 budget.
133
00:08:21.240 --> 00:08:25.080
And here's what happened behind. Again, here we are showing the logs,
134
00:08:25.200 --> 00:08:29.040
because this is important piece that we all need to understand is,
135
00:08:29.400 --> 00:08:31.520
when we give an agent a task to do,
136
00:08:32.320 --> 00:08:34.880
we need to know what the agent did with that task.
137
00:08:35.120 --> 00:08:38.720
We need to make sure that the agent understood the intent.
138
00:08:39.400 --> 00:08:44.240
As you can see here, everything that the buyer agent did is logged into logs.
139
00:08:44.400 --> 00:08:48.880
That is where you can have observability on it.
140
00:08:49.080 --> 00:08:53.240
You can evaluate it and make sure the buyer agent is doing exactly what they are
141
00:08:53.360 --> 00:08:55.200
doing. When I talk about a buyer agent,
142
00:08:55.280 --> 00:08:59.280
this buyer agent doesn't necessarily have to be created by your consumers.
143
00:08:59.520 --> 00:09:02.840
It can be an agent that you expose on your merchant website,
144
00:09:02.960 --> 00:09:07.120
too.
So don't think of the buyer agent as something that the consumers have to
145
00:09:07.120 --> 00:09:09.280
worry about. This is something that you can expose as well.
146
00:09:09.520 --> 00:09:10.760
So in this particular case,
147
00:09:11.760 --> 00:09:15.640
it is logging all of the things that Danny said about discovering those three
148
00:09:16.480 --> 00:09:17.760
providers in parallel.
149
00:09:18.680 --> 00:09:23.400
And then when robots.txt didn't allow it to go further in for
150
00:09:23.560 --> 00:09:26.200
seller one, it didn't. For seller two, it got it.
151
00:09:26.440 --> 00:09:29.560
Then it looked at the manifestation file, it looked at the capabilities,
152
00:09:29.880 --> 00:09:33.480
the capabilities matched the task, tried to execute the API,
153
00:09:33.800 --> 00:09:36.480
and it got rejected because of the 403 that we talked about.
154
00:09:36.880 --> 00:09:41.400
But the seller three was able to capture that. So if I want to visualize this
155
00:09:45.920 --> 00:09:48.400
in a more picturist way, I mean, again,
156
00:09:48.760 --> 00:09:51.880
this is the main components that any buyer agent should have.
157
00:09:52.600 --> 00:09:55.040
It should understand what the consumer intent was.
158
00:09:56.120 --> 00:10:01.000
It should discover what services that it can use to
159
00:10:01.280 --> 00:10:06.180
fulfill the task that we have given it.
And it does that using the
160
00:10:06.220 --> 00:10:09.000
manifest file that I talked about. And I just gave you a high...
161
00:10:09.460 --> 00:10:13.540
The manifest file talks about who is this, who's the seller,
162
00:10:14.140 --> 00:10:16.800
and where to send the request, how much it costs. Basically,
163
00:10:16.860 --> 00:10:19.960
it gives all of the information for the seller agent to able to act on it.
164
00:10:20.520 --> 00:10:24.440
And then it enforces the policy. So when the seller agent says, "Hey,
165
00:10:24.500 --> 00:10:29.460
it could cost me 12¢ or $2." It knows how much we told it that it
166
00:10:29.480 --> 00:10:32.820
can spend. If it is 12¢, it's under the budget.
167
00:10:32.880 --> 00:10:35.560
So it can actually go ahead and do it.
168
00:10:35.800 --> 00:10:40.260
If it is about $2 that Danny said, it would come back to Danny and say, "Hey,
169
00:10:40.400 --> 00:10:45.040
Danny, I'm not able to fulfill your task, but I can do it for $2.5 cents.
170
00:10:45.140 --> 00:10:48.360
Are you okay with it?" And as soon as Danny says "Approved," it'll go and
171
00:10:48.780 --> 00:10:52.380
execute on that.
So that's the enforcement button over there.
172
00:10:52.700 --> 00:10:57.280
And then eventually being able to pay with the wallet and then finally
173
00:10:57.320 --> 00:11:00.360
confirming the order back, confirming that the task is completed.
174
00:11:00.740 --> 00:11:05.060
So these are the main things I know: autonomously act on your user's behalf,
175
00:11:06.140 --> 00:11:10.540
raise it to consumers if there is an escalation that is needed and
176
00:11:10.840 --> 00:11:11.620
audit.
177
00:11:11.620 --> 00:11:16.040
Have everything that the buyer agent does in auditable
178
00:11:16.600 --> 00:11:19.840
file. So this is how a buyer agent works,
179
00:11:19.900 --> 00:11:23.860
but we need to look at it from a seller perspective as well,
180
00:11:23.920 --> 00:11:26.120
because if the buyer agent is coming to merchants,
181
00:11:26.600 --> 00:11:30.500
if the merchants are still using APIs or your products is not ready to be
182
00:11:31.360 --> 00:11:35.140
accessed by a agent, then you still are not going to capture the sale. So,
183
00:11:35.920 --> 00:11:38.060
Danny, can you walk us through the seller agents?
184
00:11:39.400 --> 00:11:43.980
Yeah. Let's kind of recap what happened during that prototypical demo.
185
00:11:44.920 --> 00:11:49.380
So seller one, what do the seller server logs look like?
186
00:11:49.560 --> 00:11:53.420
They're blank because I got blocked at the front door. So again,
187
00:11:53.760 --> 00:11:56.120
no visibility at all to this being a buyer.
188
00:11:59.540 --> 00:12:02.180
Seller two, a little bit better, right? We're visible.
189
00:12:02.840 --> 00:12:06.440
We got API access to the product catalog. We found a manifest.
190
00:12:07.280 --> 00:12:08.940
The payment rails just aren't configured.
191
00:12:09.040 --> 00:12:13.000
So you can see in detail in the logs that, hey,
192
00:12:13.220 --> 00:12:16.400
sign up for an account. As I said earlier, agent to agent,
193
00:12:16.720 --> 00:12:18.060
an agent's not going to sign up to your account.
194
00:12:18.280 --> 00:12:21.980
It's going to go on to seller three. So seller three,
195
00:12:22.480 --> 00:12:26.280
sale actually captured. Website discovered,
196
00:12:26.920 --> 00:12:29.900
manifest found, API catalog access granted.
197
00:12:30.540 --> 00:12:33.040
We're well within our policy to make this transaction.
198
00:12:33.840 --> 00:12:37.140
The seller agent is actually doing some things that are pretty new.
199
00:12:37.440 --> 00:12:40.760
It's a new paradigm that we're talking about.
We're checking things like,
200
00:12:41.580 --> 00:12:44.320
is this an approved transaction?
201
00:12:44.940 --> 00:12:46.600
Is this an approved catalog item?
202
00:12:46.700 --> 00:12:51.200
Photography is an approved method that we're going to implement via this
203
00:12:51.320 --> 00:12:55.360
API. Do we have inventory? Are we protecting our margin?
204
00:12:55.420 --> 00:12:57.000
There's a lot of things that we're going to talk about,
205
00:12:57.200 --> 00:13:01.960
but basically in this scenario, the seller agent met all the criteria,
206
00:13:02.400 --> 00:13:04.540
especially having the payment rail in place,
207
00:13:05.460 --> 00:13:06.880
and that was a successful transaction.
208
00:13:07.680 --> 00:13:12.040
So what does a seller agent
209
00:13:12.960 --> 00:13:15.860
actually look like? How does it work?
210
00:13:17.000 --> 00:13:20.600
It's got a lot of responsibility. And this is, again, a new paradigm.
211
00:13:20.980 --> 00:13:24.000
There were some people that raised their hands earlier that I think have started
212
00:13:24.040 --> 00:13:25.520
to implement this as a concept.
213
00:13:26.260 --> 00:13:29.660
And we've talked a lot this week about agent to agent with machine payments that
214
00:13:29.720 --> 00:13:34.040
Stripe has released, and that's a protocol that allows this to happen.
215
00:13:35.060 --> 00:13:40.060
But you do have to have some logic in place as you're going to scale this to the
216
00:13:40.080 --> 00:13:44.820
enterprise.
As you start moving away from microtransactions to real physical
217
00:13:44.880 --> 00:13:48.660
goods that have real value and a lot of monies exchanging hands here,
218
00:13:48.720 --> 00:13:53.620
but even between agents. So this is a concept where you've got to, first of all,
219
00:13:53.900 --> 00:13:56.320
authenticate the agent that's coming through the front door,
220
00:13:56.940 --> 00:14:00.080
make sure that published manifest exists.
221
00:14:00.340 --> 00:14:05.320
So that's how website two and three or seller two and three were discovered
222
00:14:06.020 --> 00:14:09.080
in the demo that I showed. That manifest is crucial to that.
223
00:14:09.860 --> 00:14:13.160
You've got to be able to quote a dynamic price.
224
00:14:13.240 --> 00:14:14.880
Pricing can change on the fly,
225
00:14:15.500 --> 00:14:19.580
especially based on things like available inventory or how much volume
226
00:14:20.540 --> 00:14:21.580
you're trying to purchase.
227
00:14:22.700 --> 00:14:26.860
That kind of goes into step four, which is negotiation.
228
00:14:27.860 --> 00:14:30.880
We have to enforce basically policies.
229
00:14:31.580 --> 00:14:34.560
So this was mentioned in the keynote earlier this week,
230
00:14:34.660 --> 00:14:39.100
but agentic commerce is powering a transaction that's becoming
231
00:14:40.640 --> 00:14:44.420
less of a moment and more of a policy.
So as policies are in place,
232
00:14:44.550 --> 00:14:46.710
Kalyani talked about buyer's agent policies.
233
00:14:46.770 --> 00:14:48.870
It's important as a consumer that you're protected.
234
00:14:49.230 --> 00:14:53.790
It's important as a seller that you've got the same level of policy engine in
235
00:14:53.830 --> 00:14:58.610
place to protect things like your margin, your inventory.
236
00:14:58.770 --> 00:15:02.990
Make sure that all of these guardrails are in place as you move forward.
237
00:15:03.430 --> 00:15:05.310
So these are just eight different categories.
238
00:15:05.890 --> 00:15:09.230
The very key category is "collect." So, collect,
239
00:15:10.230 --> 00:15:15.150
think of it as this: agents reason, but code actually pays.
240
00:15:16.110 --> 00:15:19.970
So the protocols are very, they're very solid.
241
00:15:20.670 --> 00:15:23.110
We've talked about them a lot this week from Machine Payments Protocol.
242
00:15:24.030 --> 00:15:26.050
AI powers a lot of the reasoning,
243
00:15:26.090 --> 00:15:28.430
is powering most of the boxes that you see there.
244
00:15:28.870 --> 00:15:31.410
But when we move into a payment rail,
245
00:15:32.030 --> 00:15:35.950
this is something that AI steps out, and it just executes a call,
246
00:15:36.390 --> 00:15:40.110
and that code is actually making the payment.
So it's a very key concept.
247
00:15:40.530 --> 00:15:42.350
And then we're also worried about things like fraud,
248
00:15:42.870 --> 00:15:45.490
fraud prevention that we're going to talk a little bit about how that's powered
249
00:15:45.530 --> 00:15:48.690
by the combination of AWS and Stripe and our tech stack.
250
00:15:48.770 --> 00:15:52.090
But just to give you an overall concept of a seller's agent, again,
251
00:15:52.150 --> 00:15:55.710
you're protecting the seller the same way the policy engine is protecting the
252
00:15:55.750 --> 00:15:56.583
buyer.
253
00:15:56.850 --> 00:16:01.750
And these are the key steps that you have to take and the key parts of
254
00:16:01.870 --> 00:16:05.010
logic that you have to implement that have a truly successful seller agent
255
00:16:05.110 --> 00:16:09.050
that's going to allow you to scale these microtransactions up to enterprise
256
00:16:09.090 --> 00:16:09.923
level.
257
00:16:13.610 --> 00:16:17.230
So I know Danny and I talked about the buyer agent, seller agent architecture.
258
00:16:17.330 --> 00:16:19.130
Again, those are sample architectures.
259
00:16:20.310 --> 00:16:25.170
Your agents could be a little bit more complex. We touched on negotiation,
260
00:16:25.310 --> 00:16:29.600
pricing. Let's say you have an item for $12,
261
00:16:30.970 --> 00:16:33.810
and that's a reasonable price for single item.
262
00:16:33.870 --> 00:16:37.890
But if I wanted to buy in bulk, if I'm going to buy 1,000 pieces of it,
263
00:16:38.750 --> 00:16:42.610
if I came to the store, or if I had a representative that I talked to,
264
00:16:43.050 --> 00:16:45.200
you will negotiate that price and give it to me for $8.
265
00:16:46.270 --> 00:16:51.270
But if that agent is sitting on your website and is able to do that with the
266
00:16:51.310 --> 00:16:52.430
buyer agent, then
267
00:16:54.470 --> 00:16:57.990
you can make sure that your margin is protected, but at the same time,
268
00:16:58.050 --> 00:17:01.790
quote a price that the buyer agent can actually take an action and buy it.
269
00:17:02.190 --> 00:17:04.670
So you are actually capturing that sale. So,
270
00:17:06.010 --> 00:17:10.210
talking about these things, the building blocks are ready,
271
00:17:10.510 --> 00:17:15.410
right?
You have AWS. AWS has several key functionalities,
272
00:17:15.690 --> 00:17:20.430
features, and products that can help you build and scale agents in
273
00:17:21.430 --> 00:17:23.650
however your business needs require.
274
00:17:24.530 --> 00:17:27.610
And Stripe is here to power the agentic payments.
275
00:17:28.960 --> 00:17:33.650
And then we also have our partners who can accelerate your journey. Again,
276
00:17:34.010 --> 00:17:38.870
there's a lot of news coming out and it's overwhelming amount of information out
277
00:17:38.910 --> 00:17:43.490
there and you're always... The most question that we get is,
278
00:17:43.910 --> 00:17:44.930
how do we even get started?
279
00:17:44.990 --> 00:17:48.330
There's so much noise or there's so much information, we can't keep up with it.
280
00:17:48.710 --> 00:17:51.530
So that is our job, our partner's job.
281
00:17:52.150 --> 00:17:55.970
We keep ourselves informed on what's happening in the industry and how to go
282
00:17:56.050 --> 00:18:00.770
about implementing some of these things, protecting you as a merchant,
283
00:18:00.830 --> 00:18:02.770
but also your consumers.
284
00:18:03.810 --> 00:18:06.250
So if you want to go from concept to production faster,
285
00:18:06.930 --> 00:18:08.170
we have partners for that. Again,
286
00:18:09.430 --> 00:18:11.920
we are just touching on three important things. We have...
287
00:18:12.850 --> 00:18:17.830
Come find us afterwards if you want to talk about any specific use cases.
So
288
00:18:17.910 --> 00:18:21.990
as Danny said, a new type of consumer is emerging,
289
00:18:22.570 --> 00:18:25.950
and for the first time in our history,
290
00:18:26.250 --> 00:18:31.090
we actually have agents surpass the traffic on
291
00:18:31.150 --> 00:18:32.930
websites than us
292
00:18:33.370 --> 00:18:36.230
humans.
293
00:18:36.230 --> 00:18:40.150
And 51% number may not look like a big number,
294
00:18:40.450 --> 00:18:43.110
but the number next to it, if you look at it, that is
295
00:18:45.220 --> 00:18:50.010
7,851% year-over-year growth in agentic AI traffic.
296
00:18:51.450 --> 00:18:56.430
So depending upon how your website or ecommerce site
297
00:18:56.470 --> 00:18:59.710
is set up, or how your infrastructure is set up,
298
00:19:00.130 --> 00:19:02.470
you may or may not be seeing this traffic.
299
00:19:03.010 --> 00:19:07.990
Just because your logs are not showing you agentic traffic doesn't mean
300
00:19:08.290 --> 00:19:11.730
the agent hasn't tried to come to your website. It did;
301
00:19:12.450 --> 00:19:13.730
you were just not aware of it.
302
00:19:14.490 --> 00:19:18.050
So I won't dwell more into the thing.
303
00:19:18.130 --> 00:19:21.230
So this is the reason why we are talking about it.
304
00:19:21.570 --> 00:19:24.710
We are still in the stage of innovation. It's not like everything,
305
00:19:24.850 --> 00:19:27.870
all the tech stack is completely built and ready to jump on,
306
00:19:28.250 --> 00:19:31.510
but we need to go on this journey together because this is happening.
307
00:19:34.330 --> 00:19:38.570
So this is my favorite slide in all of the slides that we have,
308
00:19:39.010 --> 00:19:42.930
because it's all some of my problems. So, where we see traction.
309
00:19:43.010 --> 00:19:46.050
So Danny and I have been having a lot of conversations with customers,
310
00:19:46.570 --> 00:19:51.450
and where we are seeing a lot of adoption of agent-to-agent commerce is in the
311
00:19:51.510 --> 00:19:52.343
SaaS.
312
00:19:52.510 --> 00:19:56.710
I know you must have heard about all of these pay-as-you-go models coming up,
313
00:19:56.850 --> 00:20:00.030
API monetization, talk about API monetization,
314
00:20:00.950 --> 00:20:05.090
monetizing your data. Some of those things used to be Herculean tasks before,
315
00:20:05.390 --> 00:20:07.170
but now it's much more easier.
316
00:20:07.610 --> 00:20:10.450
So you can capture the cash and carry
317
00:20:12.550 --> 00:20:14.190
type of a thing by automating,
318
00:20:14.610 --> 00:20:18.930
accepting payments using the protocols that Danny was talking about.
319
00:20:20.830 --> 00:20:22.750
Automated vendor evaluation, I don't know,
320
00:20:23.510 --> 00:20:26.330
everybody needs a vendor software,
321
00:20:26.390 --> 00:20:30.550
a piece of software that you want to use within your infrastructure,
322
00:20:30.610 --> 00:20:35.270
but sometimes that evaluations take six months.
But if you have an
323
00:20:35.370 --> 00:20:39.570
agent that can evaluate two or three services at the same time in parallel and
324
00:20:39.650 --> 00:20:43.850
come back to a recommendation, then your time to market increases, right?
325
00:20:44.130 --> 00:20:46.730
So the sooner you can get to your consumers,
326
00:20:46.830 --> 00:20:50.690
the sooner you can capture that revenue. So we are seeing a lot of that.
327
00:20:50.770 --> 00:20:53.410
And then I also talked about dynamic pricing.
328
00:20:53.830 --> 00:20:57.490
So a lot of ecommerce websites have pricing that is available,
329
00:20:58.070 --> 00:21:02.910
but that pricing is sometimes based on the inventory and
330
00:21:02.930 --> 00:21:05.550
everything else. You have some discounts that you offer,
331
00:21:05.650 --> 00:21:07.530
some coupons that you offer,
332
00:21:07.610 --> 00:21:11.990
but there is no mechanism where you can adapt that pricing to the needs of the
333
00:21:12.030 --> 00:21:14.710
customer. If customer is trying to buy in bulk, as I said,
334
00:21:15.710 --> 00:21:16.630
you're not able to do that.
335
00:21:16.690 --> 00:21:20.970
And now with agentic transformation that's happening,
336
00:21:21.610 --> 00:21:24.550
you have all the tools to be able to do some of the dynamic pricing
337
00:21:26.110 --> 00:21:29.470
thing.
In the travel and hospitality, field,
338
00:21:32.200 --> 00:21:34.670
I may or may not use-personally,
339
00:21:34.850 --> 00:21:38.050
I may or may not use a bot to book my travel,
340
00:21:38.770 --> 00:21:43.770
but I definitely could use a bot if it has any rebooking of
341
00:21:43.870 --> 00:21:46.370
flights necessary or hotels in seconds.
342
00:21:47.150 --> 00:21:50.870
Is there anybody in the audience whose travel was disrupted because of a
343
00:21:50.950 --> 00:21:51.783
snowstorm,
344
00:21:52.770 --> 00:21:57.650
a flight canceled?
I'm sure everybody.
345
00:21:58.030 --> 00:22:02.230
So in fact, I'll tell you a story. A couple of months back,
346
00:22:02.390 --> 00:22:06.450
I was hosting a co-innovation work stream, and I was preparing for it.
347
00:22:07.350 --> 00:22:12.070
I live in New York, and it's February, huge snowstorm, flight got canceled.
348
00:22:13.370 --> 00:22:18.350
So I had to go. As I'm preparing for my important session with my customer,
349
00:22:18.470 --> 00:22:21.950
I had to stop what I was doing, context switch to my travel agency,
350
00:22:22.770 --> 00:22:27.070
cancel my hotel because my flight was rebooked to a four days later,
351
00:22:27.130 --> 00:22:29.750
which I can't go because that's not useful for me anymore.
352
00:22:30.210 --> 00:22:34.170
So I had to go cancel my flight-cancel my flight, cancel my hotel, everything,
353
00:22:34.670 --> 00:22:36.090
come back and start working.
354
00:22:36.290 --> 00:22:40.030
But a part of me wanted to be there in person because this was an important
355
00:22:40.170 --> 00:22:42.870
innovation work stream that I was running. I wanted to be in person.
356
00:22:42.970 --> 00:22:47.810
Everybody else was going to be in person.
So like anybody
357
00:22:47.910 --> 00:22:50.190
else, me and my colleague, what we did,
358
00:22:50.290 --> 00:22:55.270
we kept looking at our travel site every few minutes to see if by any
359
00:22:55.410 --> 00:22:59.210
chance there is a seat that will open up. And it did.
360
00:22:59.670 --> 00:23:03.430
We were able to book, then we booked the hotel,
361
00:23:03.810 --> 00:23:07.230
and then we travel, and that session was successful.
362
00:23:07.290 --> 00:23:11.030
But I couldn't have imagined making that session happen remotely.
363
00:23:12.230 --> 00:23:16.170
And that's my personal experience. But throughout that journey, I was thinking,
364
00:23:16.570 --> 00:23:20.910
how I wish there was an agent that the travel agency could give us where
365
00:23:21.670 --> 00:23:25.590
it knows my policies, company's policies, it knows where I'm traveling,
366
00:23:25.650 --> 00:23:28.170
it knows my constraints as to when I need to reach there.
367
00:23:28.710 --> 00:23:32.750
It would just book it so that I don't have to do context switching. As it is,
368
00:23:32.810 --> 00:23:34.930
we are distracted with so many things that are happening.
369
00:23:35.230 --> 00:23:38.580
This is one more thing that was added to my list. So I wish...
370
00:23:39.370 --> 00:23:41.650
So that's why travel and hospitality disruption recovery,
371
00:23:41.710 --> 00:23:45.270
I would love for anybody to please build that agent. I would use it.
372
00:23:47.350 --> 00:23:51.910
And sometimes it's a multiparty trip planning, right? And if I have
373
00:23:53.760 --> 00:23:56.930
a large number of family members traveling from one location to another
374
00:23:56.970 --> 00:23:57.803
location,
375
00:23:57.890 --> 00:24:01.890
I wish I don't have to call an agent and talk to them and negotiate the price.
376
00:24:02.030 --> 00:24:04.770
I wish there was an easier way to do some of those things.
377
00:24:04.830 --> 00:24:06.110
So we see some of this.
378
00:24:06.230 --> 00:24:10.490
So we see some use cases in that travel and hospitality industry,
379
00:24:10.530 --> 00:24:11.470
that where we can see it.
380
00:24:12.390 --> 00:24:17.130
Retail and ecommerce and all industries have some B2B use cases or
381
00:24:17.190 --> 00:24:20.910
not. Again, I won't go into each of these use cases.
382
00:24:20.970 --> 00:24:23.890
These are just representatives. There's many, many,
383
00:24:24.030 --> 00:24:27.450
many more use cases where this could be useful. So
384
00:24:29.870 --> 00:24:34.230
moving along, if you want to build any of those use cases,
385
00:24:34.470 --> 00:24:36.830
AWS AI portfolio, I just wanted to show it over here,
386
00:24:38.110 --> 00:24:41.310
the entire tech stack of point in time,
387
00:24:41.730 --> 00:24:44.730
because AWS is constantly releasing new features,
388
00:24:44.790 --> 00:24:49.510
new capabilities based on your feedback.
So this is what we currently have.
389
00:24:50.050 --> 00:24:50.890
I may, again,
390
00:24:51.230 --> 00:24:54.890
things may have changed from the time I've prepared this slide to presenting
391
00:24:54.910 --> 00:24:56.790
this slide, too. So
392
00:24:59.630 --> 00:25:04.510
AWS has all the AI portfolio that you need to build and run the agents at scale,
393
00:25:05.010 --> 00:25:08.170
but then, Danny, can you walk us through what Stripe has?
394
00:25:10.290 --> 00:25:14.610
Yeah, absolutely. So kind of to recap exactly what Kalyani
395
00:25:14.830 --> 00:25:16.490
just said,
396
00:25:16.690 --> 00:25:20.050
AWS has the power to build the agents on both sides of that equation.
397
00:25:21.390 --> 00:25:25.770
Stripe is that layer of payments and that layer of trust that helps make this
398
00:25:25.970 --> 00:25:30.250
agent-to-agent transaction an actual reality. So,
399
00:25:31.110 --> 00:25:35.390
what have we launched together with AWS? So Stripe's powered by AWS.
400
00:25:35.590 --> 00:25:37.430
We have a strong partnership with AWS,
401
00:25:38.170 --> 00:25:42.270
and our go-to-market strategy and in agentic commerce has been,
402
00:25:42.910 --> 00:25:47.370
"Let's provide as many turnkey value-added solutions to our customers as
403
00:25:47.430 --> 00:25:51.030
possible." So one you've probably heard a lot about is Agentic Commerce Suite,
404
00:25:51.350 --> 00:25:55.890
which is our turnkey solution that allows you to sell your products in a
405
00:25:56.130 --> 00:26:00.750
one-to-many integration shape to many different AI companies
406
00:26:01.130 --> 00:26:05.710
and AI agents. So you've heard exciting news about the Meta launch, Gemini,
407
00:26:06.170 --> 00:26:08.690
many others are coming. So basically,
408
00:26:08.930 --> 00:26:11.890
you definitely want to start to participate as we move into the enterprise scale
409
00:26:12.590 --> 00:26:16.910
of integrating with Agentic Commerce Suite.
Stripe machine payments that powered
410
00:26:16.930 --> 00:26:21.330
a lot of the x402 money movement that we saw in the prototype?
411
00:26:22.150 --> 00:26:23.750
This is a live protocol today.
412
00:26:24.170 --> 00:26:27.310
There's about 90 companies that have already launched.
413
00:26:27.610 --> 00:26:31.590
This is the ability to monetize your API layers and different things on the
414
00:26:31.630 --> 00:26:36.530
seller side with a Machine Payment Protocol to make that a very safe and secure
415
00:26:37.170 --> 00:26:38.950
experience. Shared Payment Tokens,
416
00:26:39.090 --> 00:26:43.670
this was an early mover in the agentic token space for Stripe.
417
00:26:44.550 --> 00:26:48.190
And think of it as context wrapping around a tokenized payment method,
418
00:26:48.450 --> 00:26:51.590
things like budgeted amount, fraud signals, built in exploration.
419
00:26:52.150 --> 00:26:56.430
So this is what makes it a very safe experience for the
420
00:26:56.550 --> 00:27:01.270
consumer to make sure that when you're checking out on like a ChatGPT
421
00:27:01.510 --> 00:27:02.343
or a Gemini,
422
00:27:02.590 --> 00:27:06.550
when you give it your credit card and you're going to buy a $5 item,
423
00:27:07.510 --> 00:27:11.810
the AI doesn't hallucinate and try to buy a $5,000 item.
Even if it does try,
424
00:27:12.170 --> 00:27:15.110
it will fail because a Shared Payment Token is in place.
425
00:27:15.670 --> 00:27:20.450
Stripe Radar is that fraud protection layer that's built into the Shared Payment
426
00:27:20.450 --> 00:27:21.290
Token functionality.
427
00:27:21.830 --> 00:27:25.730
This is powered by something called the Payments Foundation Model,
428
00:27:25.890 --> 00:27:28.510
which we co-developed and launched with AWS.
429
00:27:28.870 --> 00:27:31.150
So we're very proud of that payments foundation model.
430
00:27:31.490 --> 00:27:35.730
So Stripe Radar is powering the fraud prevention layer in agentic commerce
431
00:27:35.870 --> 00:27:36.703
today.
432
00:27:36.850 --> 00:27:41.490
Link's wallet for agents is something that we launched this week as part of
433
00:27:41.530 --> 00:27:42.363
Sessions.
434
00:27:42.950 --> 00:27:46.730
If you've built your own general purpose agent using something like OpenClaw or
435
00:27:46.750 --> 00:27:47.650
anything else like that,
436
00:27:47.750 --> 00:27:52.690
you can actually install a skill called Link CLI that will allow your agent
437
00:27:53.170 --> 00:27:57.910
to go make purchases on the internet with those guardrails in place.
438
00:27:57.970 --> 00:27:59.290
So the guardrails are very important.
439
00:27:59.870 --> 00:28:02.270
And if it somehow goes outside of the guardrails,
440
00:28:02.330 --> 00:28:05.750
it will prompt you so there's human in the loop.
There's a lot of really cool
441
00:28:05.830 --> 00:28:10.110
demos. I built a demo on Link wallet for agents at the Stripe agentic booth,
442
00:28:11.070 --> 00:28:15.390
so please come by. And finally, let's not forget about MCP,
443
00:28:16.350 --> 00:28:17.650
the protocol from the past.
444
00:28:18.250 --> 00:28:22.690
Stripe can definitely allow you to monetize your MCP apps by
445
00:28:22.830 --> 00:28:26.890
using our MCP server and our MCP toolkit. So
446
00:28:29.410 --> 00:28:31.850
quickly quick, the roadmap, get visibility.
447
00:28:32.010 --> 00:28:36.470
Do not put that brick wall in front of your website so the agents get blocked at
448
00:28:36.510 --> 00:28:40.790
the front door. Second, monetize your... I'm sorry,
449
00:28:41.030 --> 00:28:44.290
modernize. See, Stripe, monetize, right? Modernize-.
450
00:28:44.450 --> 00:28:45.130
You should monetize too.
451
00:28:45.130 --> 00:28:50.010
Your tech stack. Okay. Make sure that you've got things like an agent manifest.
452
00:28:50.070 --> 00:28:53.950
Make sure that you've got a programmable API-accessible catalog.
453
00:28:54.310 --> 00:28:56.170
You have to basically modernize.
454
00:28:57.570 --> 00:29:00.710
Leverage turnkey solutions like Agentic Commerce Suite and some of the other
455
00:29:00.730 --> 00:29:03.690
solutions we've discussed. Build focused agents.
456
00:29:04.010 --> 00:29:07.850
So you basically want to start small, build an agent that has smaller tasks,
457
00:29:08.410 --> 00:29:10.850
and then move in time into other use cases.
458
00:29:11.450 --> 00:29:15.910
Design for fraud and trust. Again, implementing things like Stripe Radar,
459
00:29:16.230 --> 00:29:20.310
Shared Payment Tokens, make it a very trustworthy experience.
460
00:29:20.830 --> 00:29:23.910
And finally, shape the standards. We want to hear from you as customers.
461
00:29:24.270 --> 00:29:27.090
These are your buyers. These are your customers at the end of the day.
462
00:29:27.650 --> 00:29:29.690
We're all in between trying to make this a good experience,
463
00:29:29.810 --> 00:29:32.450
but we want you to help us shape the standards as we move forward.
464
00:29:33.570 --> 00:29:34.850
Thank you all for joining us today.