00:00:01.300 –> 00:00:06.920
Welcome to the Jason and scot Show, your source for the latest news and trends

00:00:06.920 –> 00:00:08.530
in the e-commerce industry.

00:00:08.640 –> 00:00:13.870
Featuring host Jason Retail Geek Goldberg, Chief Commerce Strategy Officer of

00:00:13.870 –> 00:00:18.960
Publicis, and scot Wingo, co-founder and CEO of ReFiBuy.

00:00:19.280 –> 00:00:21.720
Here are Jason and scot.

00:00:22.540 –> 00:00:27.320
Hey everybody, it is scot Wingo, the better half of the Jason and scot Show.

00:00:28.040 –> 00:00:32.409
In the last year, we have experienced a rapid rise of agendic commerce,

00:00:32.409 –> 00:00:38.070
this disrupted discovery, and I believe ultimately will work its way into the whole purchase funnel.

00:00:39.449 –> 00:00:43.149
There’s been lots of innovations in agendic commerce, like the universal cart

00:00:43.149 –> 00:00:47.929
that just hit general availability out of Google Gemini, dynamically generated

00:00:47.929 –> 00:00:52.619
user experiences from ChatGPT just recently with GPT-6’s release.

00:00:53.119 –> 00:00:58.459
And in the last 40 days, new personal shopping agents like Muse and Instinct

00:00:58.459 –> 00:01:03.219
that are really capturing consumers’ behavior and wallet.

00:01:04.799 –> 00:01:11.309
We’ve also had Chachpti start from zero and scale an ad business to a billion-dollar

00:01:11.309 –> 00:01:14.369
run rate faster than we have ever seen before.

00:01:15.119 –> 00:01:19.779
We also know that Amazon and Walmart are both exploring what advertising innovations

00:01:20.169 –> 00:01:24.939
around adjusted commerce, like sponsored prompts, would look like in their retailer

00:01:24.939 –> 00:01:28.559
shopping agents like Amazon AFS, previously known as Rufus,

00:01:29.009 –> 00:01:30.719
and Walmart Sparky.

00:01:32.579 –> 00:01:38.499
This intersection of agentic commerce, AI, the changing discovery landscape,

00:01:39.239 –> 00:01:43.179
and what that all means for traditional advertising models like retail media

00:01:43.179 –> 00:01:48.009
networks, and where this is all going, is a very interesting topic to every

00:01:48.009 –> 00:01:50.129
retailer and brand that I talk to.

00:01:50.269 –> 00:01:53.869
It’s also really interesting to us here as well.

00:01:55.389 –> 00:01:59.069
We think it’s important enough and media enough to justify its own podcast.

00:01:59.069 –> 00:02:03.729
So today, we are excited to announce that we have partnered with none other

00:02:04.029 –> 00:02:09.250
than Curie Masters to launch a new podcast called AdGentic.

00:02:09.500 –> 00:02:11.750
A-D-G-E-N-T-I-C.

00:02:12.910 –> 00:02:13.460
AdGentic.

00:02:14.760 –> 00:02:18.090
AdGentic focuses on advertising in the age of AI.

00:02:18.710 –> 00:02:22.930
The first episode is out, and we thought you would enjoy checking it out because

00:02:22.930 –> 00:02:28.659
it is very similar to what a lot of topics we touch on lightly here at the Jason scot Show.

00:02:29.589 –> 00:02:32.369
If the topic is of interest to you and you enjoyed this first episode,

00:02:32.369 –> 00:02:35.829
be sure to subscribe to the audio in your favorite podcast app,

00:02:36.229 –> 00:02:40.329
the video on YouTube, and we also have a companion substack if you want to be

00:02:40.329 –> 00:02:45.020
notified when new things are put out, and it will have its own separate content as well.

00:02:45.740 –> 00:02:52.240
You can get to all of this through our handy URL, which is adgentic.ai. Again, adgentic.ai.

00:02:53.110 –> 00:02:57.650
We hope you enjoy the first episode of Adgentic, and thanks for listening to

00:02:57.650 –> 00:02:58.800
the Jason and scot Show.

00:02:59.370 –> 00:03:06.510
Happy Adgentic Commercing. Look at this. A personal AI agent doing my busy work.

00:03:07.000 –> 00:03:10.450
No banners, no sponsored results, nothing for sale.

00:03:11.050 –> 00:03:14.260
Enjoy it, because it won’t last.

00:03:14.870 –> 00:03:21.730
Ads are coming to personal AI agents. And in this video, I’ll show you why and

00:03:21.730 –> 00:03:23.420
what those ads could look like.

00:03:32.180 –> 00:03:38.030
Welcome to Adgentic, the series about advertising in the Adgentic era.

00:03:38.430 –> 00:03:45.730
I’m Kiri Masters, someone who has spent the last 10 years analyzing the retail media ecosystem.

00:03:46.010 –> 00:03:53.610
But now I’m captivated by this whole new era of advertising, ads in AI.

00:03:54.340 –> 00:03:59.800
Let’s go. So like everyone else, it seems I recently got caught up in the euphoria

00:04:00.150 –> 00:04:05.220
of setting up my own personal AI agent and letting it loose on my busy work,

00:04:05.700 –> 00:04:09.250
email, scheduling, the usual stuff.

00:04:09.620 –> 00:04:15.480
And it was magical and at first I put that down to the thrill of getting these

00:04:15.480 –> 00:04:18.990
chores off of my plate but then it dawned on me,

00:04:19.560 –> 00:04:26.510
just how clean it all really was how dutiful the AI was it only seemed to have

00:04:26.820 –> 00:04:30.080
my best interests in mind and then I clocked it,

00:04:30.839 –> 00:04:35.049
part of that feeling came from what was missing, ads.

00:04:35.989 –> 00:04:42.749
No clutter, no distractions, no reminder that somebody somewhere wants to sell me something.

00:04:42.869 –> 00:04:47.499
And I’m not the only one feeling it. Agents like Meta’s Muse,

00:04:47.879 –> 00:04:53.619
OpenAI’s Dots, and a startup called Instinct have all launched recently to a

00:04:53.619 –> 00:04:56.279
lot of love. But here’s the thing.

00:04:56.809 –> 00:05:01.319
Every big consumer platform started out exactly like this.

00:05:01.639 –> 00:05:06.659
Let’s start with Google. In 1998, its founders, Sergei and Larry,

00:05:06.939 –> 00:05:11.799
wrote the academic paper describing their search engine, and in an appendix,

00:05:12.039 –> 00:05:15.799
they warned that ad-funded search engines would be,

00:05:16.259 –> 00:05:20.299
quote, inherently biased towards the advertisers.

00:05:20.979 –> 00:05:26.139
Two years later, Google launched AdWords, and at first, the ad sat off to the

00:05:26.139 –> 00:05:28.469
side, clearly separate. it.

00:05:28.889 –> 00:05:34.449
Today, of course, they sit at the top of the page with a tiny little sponsored label.

00:05:35.019 –> 00:05:40.489
Next up, Facebook. Now, Facebook skipped the no ads promise and went straight

00:05:40.739 –> 00:05:42.729
into treating ads as inevitable.

00:05:43.350 –> 00:05:49.220
Mark Zuckerberg on 60 Minutes back in 2008 said, there have to be ads either

00:05:49.220 –> 00:05:51.040
way, because we have to make money.

00:05:51.610 –> 00:05:55.860
He even pitched them as something of an upgrade for the consumer experience.

00:05:56.490 –> 00:06:00.470
Saying, what would you rather see, a banner ad from Bloomingdale’s or that one

00:06:00.470 –> 00:06:02.640
of your friends bought a scarf?

00:06:03.050 –> 00:06:06.250
Now, remember that scarf. We’ll come back to it. And then there’s WhatsApp.

00:06:06.600 –> 00:06:12.260
Co-founder Jan Coom kept this note on his desk from his co-founder,

00:06:12.400 –> 00:06:20.040
Brian Acton. Now, Facebook bought WhatsApp in 2014, and both founders were gone within four years.

00:06:20.210 –> 00:06:24.000
And in 2025, ads arrived in WhatsApp.

00:06:24.560 –> 00:06:29.330
Amazon is a slightly different story because it started as a store rather than

00:06:29.510 –> 00:06:31.610
a publisher or a media network.

00:06:31.970 –> 00:06:37.610
But in 2025, Amazon made almost $70 billion from its ad business.

00:06:37.610 –> 00:06:42.550
That is about double the profit from all of its web stores combined.

00:06:42.850 –> 00:06:46.260
The store is the content, the ads pay the bills.

00:06:46.710 –> 00:06:55.920
And AI is already following the script. In 2024, Sam Altman called ads a last resort for OpenAI.

00:06:56.210 –> 00:07:01.500
But by early 2026, ChatGPT was showing ads to its free users.

00:07:01.940 –> 00:07:04.480
So that’s the pattern, and it has a name.

00:07:05.040 –> 00:07:10.140
It’s called enshittification. And yes, that is a swear word,

00:07:10.140 –> 00:07:14.110
but it is also in the dictionary. so it is a real word too.

00:07:14.270 –> 00:07:18.800
The writer Corey Doctorow first coined it back in 2022.

00:07:19.000 –> 00:07:23.330
His version goes, first, they are good to their users.

00:07:23.690 –> 00:07:30.600
And then step by step, the platform shifts the value to its advertisers and then to itself.

00:07:30.760 –> 00:07:36.330
And so the word stuck, it was named word of the year by the American Dialect

00:07:36.330 –> 00:07:37.970
Society. And consumers stuck.

00:07:38.590 –> 00:07:41.880
Resent in shittification. Of course they do.

00:07:42.170 –> 00:07:46.560
But here’s where I might lose some of you. Despite all of that,

00:07:47.040 –> 00:07:54.060
advertising is still the most elegant way we have to pay for consumer technology.

00:07:54.190 –> 00:07:59.300
Now, I am an advertising maximalist and I’m with Eric Sufort on this one.

00:07:59.610 –> 00:08:05.820
He says, digital advertising is the most positive sum economically expansionary

00:08:05.820 –> 00:08:10.790
technology summoned by humanity. So let’s break it down. There are three main

00:08:10.790 –> 00:08:13.710
ways to pay for a consumer app.

00:08:14.010 –> 00:08:16.280
The first is subscriptions.

00:08:16.820 –> 00:08:20.550
There’s a flat price, whether you use it once a month or all day.

00:08:21.230 –> 00:08:28.540
Second is fees, a cut taken every time money changes hands, like the service

00:08:28.540 –> 00:08:30.650
fee on your food delivery.

00:08:31.310 –> 00:08:37.760
Three is ads. Now someone else pays because they want your attention.

00:08:38.280 –> 00:08:42.810
And look at how most consumer apps actually make money today.

00:08:43.420 –> 00:08:47.890
Mobile games, subscriptions, and or ads.

00:08:48.410 –> 00:08:52.740
Streaming TV, subscriptions and or ads.

00:08:53.160 –> 00:08:56.450
Booking.com, fees and ads.

00:08:56.980 –> 00:09:00.140
DoorDash, fees and ads.

00:09:00.840 –> 00:09:05.190
E-commerce, increasingly ads. Social media, ads.

00:09:05.190 –> 00:09:10.540
Now, different types of users have different relationships with advertising and subscriptions.

00:09:10.540 –> 00:09:15.710
Subscriptions really work when a business is paying for that subscription,

00:09:16.020 –> 00:09:20.240
but consumers paying their own way are a different animal.

00:09:20.240 –> 00:09:26.730
So, how are today’s consumer AI agents planning to make money?

00:09:27.030 –> 00:09:32.370
Well, Meta says that Muse will be free for a generous allowance and over time

00:09:32.580 –> 00:09:37.150
will profit by taking a small fee from transactions.

00:09:37.150 –> 00:09:43.920
OpenAI launched Dots for paying subscribers only, but Sam Altman also said you

00:09:43.920 –> 00:09:49.200
should of course expect us to do a mass market product for billions of people someday.

00:09:49.200 –> 00:09:52.590
And OpenAI already runs ads in ChatGPT.

00:09:53.010 –> 00:09:58.490
And Instinct, while it’s free, it has raised a billion dollars and it hasn’t

00:09:58.490 –> 00:10:04.190
said anything about how it will make money. I guess they’re planning on a wing and a prayer.

00:10:04.370 –> 00:10:10.340
Now, here is my problem with subscriptions and fees, especially for AI agents.

00:10:10.530 –> 00:10:12.350
It comes down to incentives.

00:10:12.920 –> 00:10:20.070
Subscriptions cap out at the people who are willing to pay, and most people won’t.

00:10:21.140 –> 00:10:28.750
When OpenAI announced ads, about 95% of ChatGPT’s weekly users were on the free plant.

00:10:29.600 –> 00:10:36.890
Fees are a sneakier trap. A fee pays the agent every time you buy something.

00:10:37.140 –> 00:10:42.750
So now the agent has a reason to nudge you to buy something specific.

00:10:42.750 –> 00:10:48.380
And once merchants are competing for those sales, they will pay to get to the front of the line.

00:10:49.010 –> 00:10:54.600
Paying for a placement in front of a shopper, that’s retail media. Now, fees.

00:10:54.940 –> 00:10:58.430
Fees have a way of turning into ads.

00:10:58.890 –> 00:11:05.050
As Casey Newton at Platformer put it, it will soon be some team’s job to massively

00:11:05.050 –> 00:11:07.740
increase the number of transactions.

00:11:08.310 –> 00:11:15.830
And notice Meta’s language about how they’re going to monetize profit from fees over time.

00:11:16.000 –> 00:11:18.470
So no one has really ruled anything

00:11:18.470 –> 00:11:23.780
out. And that’s why I believe ads are coming to personal AI agents.

00:11:24.040 –> 00:11:27.280
The real question is, how shitty will it be?

00:11:27.680 –> 00:11:30.720
And this is where my eternal optimism kicks in.

00:11:30.960 –> 00:11:35.550
Maybe this time we don’t need to repeat the mistakes of the past.

00:11:35.550 –> 00:11:38.250
Think about why ads annoy you.

00:11:38.680 –> 00:11:41.080
Usually it’s one of two things.

00:11:41.530 –> 00:11:45.200
It’s not relevant to me or it’s getting in my way.

00:11:45.830 –> 00:11:47.310
Now remember the scarf.

00:11:48.000 –> 00:11:53.510
Plenty of us will begrudgingly admit that the ads on Instagram and TikTok are

00:11:53.510 –> 00:11:58.490
often, okay, sometimes actually helpful.

00:11:59.429 –> 00:12:05.379
AI agent that really knows you has an even better shot at getting both of those things right.

00:12:05.709 –> 00:12:09.279
The right ad at the right moment in the conversation.

00:12:09.709 –> 00:12:14.539
So far though, ads in AI have been pretty uninspiring.

00:12:15.059 –> 00:12:21.379
ChatGPT’s ads are a labelled box under the answer. Logo, headline, image, link.

00:12:21.949 –> 00:12:28.239
It’s a search ad with a fresh coat of paint. We are starting to see some new

00:12:28.239 –> 00:12:31.439
ad formats developing that are native to AI.

00:12:31.779 –> 00:12:35.399
One is sponsored prompts, which Amazon pioneered.

00:12:35.789 –> 00:12:44.079
Another is sponsored agents, where users can have a conversation with an advertiser’s AI agent.

00:12:44.369 –> 00:12:49.589
And finally, OpenAI is testing its first visual ad format in ChatGPT,

00:12:49.599 –> 00:12:53.529
which appears while someone is waiting for an image to generate.

00:12:53.979 –> 00:13:00.089
But my prediction is that agents will more likely borrow from the formats that

00:13:00.089 –> 00:13:05.819
we already know. From retail media, sponsored picks in your agent’s shortlist

00:13:05.819 –> 00:13:08.339
and offers on your order confirmation.

00:13:08.879 –> 00:13:15.359
From gaming, rewarded ads. Watch an ad, earn more free usage.

00:13:15.959 –> 00:13:22.669
From streaming TV and ad on the working on it screen while your agent books your flight.

00:13:22.889 –> 00:13:28.639
And here is the neatest one, not actually showing any ads at all.

00:13:29.299 –> 00:13:34.869
Offsite advertising is where a retailer or another company with lots of user

00:13:34.869 –> 00:13:40.489
information allows advertisers to target their users on another platform like

00:13:40.489 –> 00:13:44.189
connect to TV or the open web or social media.

00:13:44.599 –> 00:13:50.399
And advertisers love it because they can target very precisely by behavior,

00:13:50.399 –> 00:13:53.259
by predicted behavior, by different customer groups.

00:13:53.729 –> 00:13:57.919
And now think of all of that behavioral insight that these chat assistants have

00:13:57.919 –> 00:14:00.829
about us. They know our hopes, our dreams, our ambitions.

00:14:01.309 –> 00:14:05.729
And with an off-site advertising program, these AI companies would never have

00:14:05.729 –> 00:14:08.769
to distract us with an ad or get in the way.

00:14:09.309 –> 00:14:13.409
Now, the only catch is the margins of this type of advertising are typically

00:14:13.409 –> 00:14:18.129
lower because you have to share that ad spend around with more partners.

00:14:18.129 –> 00:14:23.319
And here’s my personal wish list, which I’m sending off to AdSanta this Christmas.

00:14:23.319 –> 00:14:25.949
One is opt-in brand drops.

00:14:26.179 –> 00:14:30.769
Let me tell my agent which brands I want to hear from when they launch something new.

00:14:30.999 –> 00:14:34.419
Two is sponsored announcements I asked for.

00:14:34.899 –> 00:14:40.499
And three, ads matched to what I’m actually shopping for. I really think that

00:14:40.499 –> 00:14:45.229
with this depth of information that our AI assistants have about us,

00:14:45.479 –> 00:14:49.309
the ads could be better than we’ve ever seen them on other platforms.

00:14:49.829 –> 00:14:54.289
Now, here’s the one that everyone’s worried about and what I think we need to

00:14:54.289 –> 00:14:59.139
be aware of and that is ads that quietly steer your agent.

00:14:59.629 –> 00:15:07.319
Paid rankings that you never see. A merchant paying to become your agent’s default choice.

00:15:07.869 –> 00:15:12.299
That’s what consumers already say that they are concerned about and that’s where

00:15:12.429 –> 00:15:15.629
inshidification is a real issue.

00:15:15.629 –> 00:15:20.099
It’s exactly the bias that Google’s founders warned about back in 1998.

00:15:20.099 –> 00:15:23.749
That’s why I think that ads that are shown to me that are relevant and clearly

00:15:23.749 –> 00:15:26.519
labeled are the best model that we’ve got.

00:15:27.109 –> 00:15:32.479
I don’t think ads are abuse. One day we might look back and miss the clean,

00:15:32.479 –> 00:15:41.229
quiet experience of these early AI agents, but of all the options, ads are the least worst.

00:15:41.229 –> 00:15:48.619
Thanks for tuning into Adgentic, the new show about advertising in consumer AI environments.

00:15:49.049 –> 00:15:54.349
Adgentic is part of the RefiBuy Media Network. I’m Kiri Masters,

00:15:54.859 –> 00:15:55.589
and I’ll see you next time.