ChatGPT Instant Checkout: Deep Dive Part III/III: Open Questions, FAQs (with Answers!) and Anti-Agentic Commerce Naysayers 🥊

We wrap the three part series by summarizing open questions, answering FAQs and addressing the naysayers.

Welcome to the last part of our three part Deep Dive on ChatGPT Instant Checkout. Here’s where we are in this series:

• Part I - User Experience, Merchant (DTC Brand or Retailer) Experience and what does this mean for Merchants (you, most of my readers) and Agentic Commerce strategically?

• Part II - Let’s get nerdy and dig into the feed and ACP specs to understand more.

• YOU ARE HERE——» Part III - Open questions, what the naysayers are saying, and FAQs we are hearing from merchants with answers where available.

Retailgentic Housekeeping Note

This week we published:

• Tuesday - The turn the tables pod with Kiri here.

• Wednesday - The big Walmart/Salesforce FLASH news here.

• Thursday - Today we’re closing out the ChatGPT Deep Dive with a bang!

Looking at the week’s activity and news, that’s 99% of what’s going on, so there will not be a weekend roundup in your inbox Sat/Sun. Don’t dismay though, because next week, we’re ALSO wrapping up our GenAI Engine Optimization Series with Part III (the big Agentic Commerce Optimization Playbook reveal!) that was delayed ~30 days due to the ChatGPT release. If you have time this weekend, go review parts I (here) and II (here) to refresh for part III.

Ok, back to your regularly scheduled substack post…

First, a Word from OpenAI…

We now have several new pieces of information since the release of ChatGPT on 9/30. First, we have this from Sam Altman: "…I think it is one of these things that is just like, as long as you let the merchant have the direct relationship with the consumer, which we do, I think it’s just better, that one’s just better for everyone, it’s a better user experience, it’s better for merchants.” If you want the whole transcript, this if from his Stratechery interview we covered here.

Second we have Michelle Fradin, ChatGPT’s product lead for ChatGPT Commerce. She gave two interviews: CNBC here and Fortune here. Having followed OpenAI a while, they don’t say a ton, but what they do say, they tend to mean, so I pay close attention to it. Here’s her detailed quotes from these: (bolding is me)

CNBC→

• “Our vision for ChatGPT — and a lot of the technology we create, but especially ChatGPT — is that it’s not just providing you information, it is also helping you get things done in the real world,”

• ChatGPT surpassed 700 million weekly active users in August, and Fradin said “a huge portion” of the questions people are asking are related to shopping and commerce in some capacity. The questions can be specific, such as asking for help finding a pair of black boots, or more broad, such as how to plan a birthday party for 12 kids. “What we wanted to do is make it even easier for users to complete their journey in ChatGPT, and for merchants and developers to convert those conversations into checkouts,” Fradin said.

Fortune→

• Still, why would OpenAI create a protocol that competitors can use, too? ChatGPT product lead Michelle Fradin framed it as a merchant-first calculation: “The primary goal we had was making something incredibly easy for the entire ecosystem—merchants and developers—to adopt,” she said. “So yes, one aspect of that is, competitors or other players in the space can adopt it, too.”

• At the end of the day, she explained, it’s “net-beneficial” if merchants have to do less work to integrate with different platforms and grow sales. “We felt like the most merchant-friendly approach was making this available to everyone,” she added, noting there was so much demand from merchants that “we needed to build something that could scale.”

• “The best way today to ensure as xa merchant that your product has the best chance of being chosen is to make sure that we have the most up-to-date and richest amount of information about your products as possible,” she said. Fradin added that merchants are eager to provide detailed “product feeds”—essentially structured catalogs of their items with rich descriptions, updated prices, and availability—so ChatGPT has the fullest context when deciding what to recommend. In some cases, she noted, brands are supplying even more detail than they publish on their own sites, in hopes of improving their chances of being surfaced.

What are they telling us?

This is super important so I’m going to boil it down to these 5 critical points they are telling us straight-up:

1. A ‘huge portion’ of ChatGPT’s 800m WAUs are using it for shopping

2. The best way to get your products to show up in ChatGPT is to send a feed and they want the richest data possible.

3. Retailers and brands are supplying even more detail than they publish on their own sites

4. They are extremely merchant friendly, as witnessed by the unusual Merchant-of-record setup.

5. They have created an open standard that can be adopted by competitors. Imagine if Amazon created an open-standard for their marketplace where anyone could plug in their own front end and then ‘connect’ to all the sellers and checkout infrastructure. You can’t! - That’s how incredibly merchant friendly this is.

This is a big opportunity with significant users looking to buy products (note to readers - this is foreshadowing and it will come up later which the naysayer section 😉).

Open Questions

It’s two weeks after the big announcement and we still have X open questions. Some are tactical/short-term, some are more medium-term and some are philosophical/theological.

Short/term (this year) tactical open questions:

• When does Shopify launch?

• Why wasn’t Shopify ready for the launch?

• Who will be next?

• How many merchants will be live before Thanksgiving?

Medium term (next year) questions:

• If I currently block GPTBot, OAI-Searchbot, can I send a feed and still get my products on ChatGPT? (I’m going to guess no because if I were implementing this, I’d have an agent ping the site and update the price and availability to make sure the product is in-stock and I have the right price, but haven’t validated this yet).

• What is ChatGPT charging merchants?

• When will we see expansion to multi-cart?

• Will ChatGPT checkout add support for loyalty programs?

• Beyond Stripe, how will other payment processors be fully supported, and what adaptations are needed for non-Stripe users?

• If they are so merchant friendly, will ChatGPT provide analytics or metrics to merchants about user interactions as they go through research→buy in ChatGPT?

• Will ChatGPT allow consumers to rate merchants to help other buyers avoid fraudulent or low quality merchants?

• Related - will they implement a guarantee like the Google Shopping ‘Google Guaranteed/Trusted Store’ program (now becoming Google Verified) program?

Long-term questions:

• Are we going to have two competing standards: ACP (ChatGPT/Shopify/Stripe) and AP2 (Google/Paypal/Affirm)?

• Will some companies work with both or will it be entirely balkanized?

• Will someone build some kind of a bridge between these two standards (seems smart and valuable)?

• Sam is opening up to Ads, but being very specific that they will be for discovery and add value, what’s that look like and when is it coming?

• Will Amazon or eBay or Walmart join one of these or are the incentives aligned for them to be part of the balkanization?

Frequently Asked Questions w/ Answers

As the leading analysts and over at ReFiBuy, vendor/enabler, in this space, we’ve handled hundreds of questions and tried to boil them down to X with answers:

Question: How do I get on ChatGPT Checkout before Holiday 2025?

Answer: If you’re on Shopify, keep reading. If you’re not, ChatGPT has been swamped with interest and your best bet to get in the program is to do everything you can to prepare. That means, create your feed, get it tested and get your checkout integration all wired up and tested. If you remember back in part 2, when we got nerdy, I pointed out the ‘Production Readiness Guide that lives here and told you to ‘do not underestimate the importance of this one’ - well, turns out it’s pretty important. This basically gives you a guide and a checklist to do your own UAT on your ACP complaint checkout. Then, after you are buttoned up, make a full-press pitch to the team at ChatGPT. Pro tip - use this opportunity to significantly expand and enhance your product catalog. I believe this is so foundationally important for merchants going forward that I started a company whose entire focus is building Agentic technology that helps with this and gets your catalog 100% Agentic Commerce ready.

Question: Ok, but If I’m on Shopify, what should I do?

Answer: Good news, your favorite blogger/podcast author has a pro tip - checkout this link 😉. 🤫 (don’t tell Tobe!!)

Question: Should I open an Etsy store or Shopify store to sell on ChatGPT if my current platform is either not going to adopt ChatGPT Checkout fast enough or is opting-out.

Answer: Depends.
If you sell 100% on marketplaces like eBay or Amazon and don’t have a website, I’d say this is definitely the ‘open a simple shopify store’ moment. If you sell handmade goods - go Etsy all others go Shopify.

If you are a a mid-tier or large retailer, it gets more complicated because you need to do the math on if the Juice is worth the Squeeze?

The Juice:

• 800m WAUs

• Likely lower take-rate than other marketplaces or paying the Google/Meta ad tax

• Get a promotion for being AI forward and innovating at your company🥇. Seriously though, every public company we talk to is under extreme pressure to ‘do something and get started’ and this will definitely check this box so when your CEO, board or public company investors search for your hero sku on chatgpt, you are in the game.

• The ChatGPT customer base skews younger and categories like beauty, apparel, home improvement, furniture, appliances, personal devices and healthcare are ahead of others in adoption. Latest data we have on this is here→

(Burke: Shopper Journey with GenAI, 8/1/25 - we covered it here)

• If you are a $100m-10b/yr retailer, I would set your expectation that this is not going to be near material this year. If you are smaller, have interesting products that just aren’t getting surfaced in the categories highlighted, this could be material.

The Squeeze:

• Managing inventory between X platforms is complicated and can lead to over-sells if you don’t gate inventory

• If you are in a bigger co, what’s your CFO going to think about this or your operations team and customer support? If you are omni-channel, don’t forget to let them know about it.

• There could be some confusion with ‘authority’ down the road if you decide to shutter the Shopify shop when your direct integration goes live.

• You’re still going to still need to have a very robust datafeed if you move to shopify - so that investment is ‘portable’ thanks to it living in the product catalog layer and not the ‘buying surface’ (pdp on your legacy site /pdp on Shopify).

• Factoring all this in, the lower ‘Squeeze’ option maybe pushing through some suboptimal solutions (a nice way to say ‘hacks’) with your existing site and then coming back and fixing those in Q1_26 when the dust settles. We’re helping many retailers think through this and implement the ‘push through’ so reach out if you want to talk about it.

At the end of the day, it’s up to every merchant to weigh the Juice and the Squeeze and decide what’s best for their business.⚖️🤔

Naysayers

I’ll start by sharing my biases. As an entrepreneur, I’m a generally optimistic person, some would say to do what we do in startup-land you need to be ridiculously optimistic. I’ll also say that thanks to the Jason and Scot show’s annual predictions, for the last 15 years, I’ve had a lot of practice predicting things in e-commerce which isn’t that hard, the timing is what’s hard.

That being said, to make sure we’re not getting over our ski’s it’s always good to look at what the Naysayers are and evaluate it and make sure we understand where they are coming from. There’s a lot to be learned and it helps build the argument and make Agentic Commerce proponents stronger debaters internally and externally. It’s also more persuasive to be able to ‘get inside’ the naysayers head, that’s the only way to flip them, if you can.

How do you find Naysayers?

I know about all of these Naysayers, because people alert me to them and say: “Scot, what about <naysayer> do you agree, you should do a rebuttal!” or “you should have them on the podcast with a referee” and battle it. Maybe some day, but analyzing their Anti Agentic arguments they have put in writing, I don’t think they’d like how it will go 😉.

With that background, it’s me vs. the Naysayers in a four on one melee! Let’s see which of us can persuade you 🫵!

The Naysayers that came out of the woodwork for the ChatGPT Instant Checkout announcement can be grouped into two buckets: The Payment Risk Naysayers and the Anti-Agentic Naysayers.

Payment Risk Naysayers

The Payment Risk naysayers are typically payments/Fintech people that when they evaluate the ACP protocol, especially into the part of the protocol where you can switch out Stripe for another payment processor. This is inside the Delegated Payment Spec (here - we covered it in part 2: nerdy, told you it was nerdy). If you want to refer back to it. In this world the PSP is the Payment Service Provider (Stripe or the third party).

But First a Word about PCI

Three letters make any ecommerce exec or technologist shake in their boots - they are PCI - Payment Card Industry - the cc industry has a security standard called PCI that protects the magic risky data of credit cards - the card number and the CVV and has very clear and specific rules about those and what and when credit cards can be transmitted ‘in the open’. I’ve built PCI compliant systems and it was incredibly complex and hard to build, test and certify. Avoid it at all costs, trust me. This juice is almost never ever worth the squeeze.

Enter PCI Compliant Vaults and Tokens

Because of the gut-wrenching pain of PCI, after we built our amazing PCI system, of course 🤦‍♂️, someone (I think Stripe), invented this concept of ‘we’ll do PCI for you!” Pass, don’t store, the user’s cc to us, we’ll store it and give you a ‘token’. For example, let’s use an easy one and call it X. The token allows you then to say, “hey, this customers credit card is X”. Then you say - “hey stripe, user with credit card X (you can see the last 4 in a lookup, but that’s it) has agreed to another charge of $20” - and stripe looks in their vault, and have stored a little db that says oh, user X, has an amex 1234 5617 8901 234 and CVV 1234 (this is a pretend number, don’t panic, please don’t call PCI!). Stripe marry’s their vault with their processing, a checkout and all kinds of little payment lego blocks to solve different things.

Ok now we’re ready to tackle these Payment naysayers…

First Up Nik Milanovic - Payment Naysayer

Here’s Nik’s post on Linkedin, which I’ve captured for you to also read in-line:

What I think Nik is missing is captured in this statement: “If Stripe owns the Buy Button”. We can stop right there because this makes no sense and it’s clear Nik is very smart at payments, but he hasn’t really looked at or seen how this works. For the Instant Checkout implementation of ACP, ChatGPT shows the buy button on the product card offer and when you click on it, you are taken to a (still!) ChatGPT hosted checkout. In the delegated payment spec, any marketplace can build their own checkout and under the hood you can have a different vault and a different PSP.

Nik is right about this a bit, but Shane explains it better.

Shane Curran - Payment Naysayer

Speaking of Vaults, here’s Shane Curran’s post on Linkedin→

Now, before we address Shane’s point, did you see where he works? Yep, he works for a vault company. Shane is right, if you choose not to go the non-stripe payment option with delegated payment, you’re going to not only need a processor but you are going to need to solve PCI. What Shane may not realize is that, let’s say you are a large merchant, or your a e-commerce software provider, you have already had to solve this, because you are running a lot of cards already and you’d know if you are not PCI compliant.

There’s really three options here:

• Use Stripe - built in vault and processor - good, PCI compliant

• No-stripe/DIYPCI - Leverage your existing PCI system and processor - still PCI compliant. Your existing PCI work folds right in here and it’s just another card.

• No-stripe/Vault+PSP - Use the non-stripe option and use a third-party vault and use a different payment processor - also PCI compliant. Our buddy Shane here would strongly recommend Evervault. But be aware that Stripe has unbundled theirs and I hear good things about VGS -details here.

Where you are NOT PCI complaint is if you go non-stripe option, and decide not to use a vault. I haven’t met a modern e-commerce company that wouldn’t immediate see the problem with that, but let’s assume they do, is ChatGPT going to send users into a PCI problematic checkout situation!?

I have a feeling the people at OpenAI/ChatGPT that wrote this pretty sophisticated Delegated Payment Spec are going to have PCI compliance pretty high on their checklist (remember there’s a certification process, where before ChatGPT sends any merchant their precious users, they are going to make darn sure it’s secure, I assure you). Let’s assume I’m wrong and ChatGPT has a bunch of payment noobs and they don’t understand this - the nice people at PCI will be on top of this lickety-split.

Payment Naysayers Scoreboard:

• Nick: Didn’t understand how the buy button works

• Shane: Trying to make people scared to use his vault

Score: ACP: 2, Naysayers: 0

Anti-Agentic Naysayers

We live in a time that, for some reason, we rarely argue about topics with logic, but we end up arguing about these really silly things that boil down to word definitions. In my first career arc, I managed developers. Developers love to argue about this constantly. The two worst ones are NULL vs. 0 and tabs vs. spaces. I get there’s a personality type for which this is either important or fun to argue or both, but we’re focused on the big picture here -I believe this stuff is existential, it’s going to change commerce and retail substantially. If you make a decision based on a word definition for an existential crisis, it will be a huge mistake.

Anti-Agentic Naysayer: Payments in Full / Andrew Dresner

On his payments oriented substack, ‘Payments in Full’ Andrew has this very Anti-Agentic Commerce piece→

Agentic Commerce: Less than meets the eye

Both of our Anti-Agentic Commerce examples make a bunch of anti-Agentic Commerce positions, I'm going to go with their top 3.

Position 1: The market for Agentic Commerce is tiny

Here he makes a very weird decision to start with the 2023 household budget from the Bureau of Labor and Stats. Turns out everyone in ecommerce knows the US Census bureau has a retail and ecommerce data report that tells us in 2024 ecommerce was $1.2T and retail was $7T. He also goes through this weird narrowing of things where he basically says you don’t use agentic for high-considered purchases, you don’t use it if you want a low price (just use Amazon which is funny as they aren’t really lowest price on many items) and you don’t use it in the discovery process. Turns out there’s two word definition things going on here. He comes to these conclusions because in his definition of Agentic Commerce, the user has zero involvement. You say - buy me a couch and it just does it! No user would want that, so that excludes all the categories where don’t already know what you want.

In my definition of Agentic Commerce, there’s a wide spectrum of use cases. Agentic Commerce is my buddy shopping agent. It helps me Research products (discovery), it helps me Find the lowest price and then, if I’m cool with it, it can buy. But I’m fine pressing the buy button, like ChatGPT, I don’t look at that, throw my remote at the TV and scream THATS NOT AGENTIC AAAAHHHGGG😱. Sure, what he describes is a destination. That’s why I have my 5-level charts.

Anyway, this definition allows him to say the TAM is $20T which is all of spend, then this matrix narrows it down to $400B which is SAM and then, because competitors have it all locked up (coming up next) he says - the opportunity is $20B - tiny!

I disagree, the GMV TAM in physical products is $7T, the GMV SAM is $600B - like half of e-commerce today, by next year with multi-cart and other features it will expand to 1.2T. In my world, all of that is a jump ball for AC so SOM=SAM.

Position 2: Legacy competition

He mentions Zillow because he’s big into a house buying example as anti-agentic - ok, I give you this one <sigh>. Then he mentions Amazon and Walmart and, you ready for this, Google Shopping (wut? does he know how Google Shopping works?).

What’s hilarious about this argument is Walmart just announced that they are going to sell on ChatGPT. Andrew hasn’t talked about that yet. When you make predictions and they go horribly wrong, it’s the best indicator your view of the World, incentives, and Commerce is 180 degrees wrong.

Position 3: Retailers are blocking agents today

He’s right. I’ve talked to maybe 100 retailers that are blocking agents and you know what? The CEO didn’t know. Someone in IT chose Cloudflare and Cloudflare decided ‘agents bad!’ and is blocking them all. But you also know what? When they find out this gets solved in about oh, 24-48hrs.

Also, if you do decide to block agents for some reason (don’t do this), you can send, are you sitting down, (whispers) a direct feed of your

In fact, the very ACP protocol talks about this in great detail. Come on Andrew, do better buddy - I knocked that out with my right hand behind my back 🥊

Anti-Agentic Naysayer: Andrew Lipsman

What is it with Andrews and hatin’ on agentic. This is the only Naysayer in the bunch that I actually know from ecommerce, so surely he’s going to have more dialed in tighter arguments. Let’s see!

Here’s his piece from October 6. This is after ChatGPT announced ACP and only 10 days ago. Can we all agree I didn’t go cherry pick something from 2024 or something here?

Position 1: Consumers don’t want Agentic Commerce

Andrew, let’s do some math here. 800m WAU’s on ChatGPT alone, put that at 1B for all the answer engines.

Andrew: “0% want agentic”

Scot: Ummm, hey, let’s ask some experts. Who would be good to ask? Let’s start with OpenAI. If only, they released actual usage data. I bet they didn’t because 0 is so embarrassing. Wait, this just in!

Economic Research Chatgpt Usage Paper (1)

9.22MB ∙ PDF file

Download

Download

If only there was a substack that covered this ☹️. Oh wait! There is? What’s this - some rando has done math on this and, guess what? It’s a bit more than 0→

Don’t believe me and OpenAI? How about Melisa’s survey of over 1000 us consumers?

How about this small bank called Morgan Stanley?

Position 2: Yep, you guessed it - word thinking.

<sigh> Andrew a very smart guy, surely he doesn’t really look around and see 0 people interested or using Agentic Commerce, what’s up here. Well, conveniently he drops his definition.

Look familiar? Yep, ever Anti-Agentic believer’s foundation is built on a very limited and almost religiously applied (spaces vs tabs!!!) fervor:

AGENTIC MUST HANDLE THE ENTIRE BUYING PROCESS!!!!

-Anti Agentic Enthusiasts screaming loudly as all the shoppers leave their site.

What’s ironic about this is Andrew is strongly advising Retail Media Networks execs: “shouldn’t waste a single moment preparing for this “threat.” and this advice is going to get these people REKT.

Here’s my logic:

• RMNs are ads on ecommerce sites. Think of all the stuff on Amazon’s home page and in the search results that keep you from finding the product you are looking for.

• All the paths to see an ad for RMN are in the Research and Find part of the funnel. Once you get to a PDP you’ve survived the gauntlet and finally can see some non-ad information (the product detail page).

• In my definition (wide spectrum of helping the user with agents), they have all started at the Research/Find and until ChatGPT checkout, you would then go to the PDP to transact.

• Guess who makes $0 in that flow? RMNs

• Also, have I ever mentioned clicks from Google are trending to zero as a reaction to Agentic Commerce? You think that is going to be good for RMNs when all that traffic to the home page goes away and instead it’s replaced by off-site transactions OR Agents sending traffic to PDPs - both of those paths result in $0 for the RMNs.

Andrew’s tight definition of Agentic Commerce has an entire sub-section of ecommerce ignoring the biggest existential crises they have ever faced.

If only there was a big event 🤔 where Andrew would snap out of this World view….

Position 3: Retailers Don’t Want Agentic Commerce

Maybe Andrew didn’t see this, but ChatGPT Checkout has two launch partners: Etsy (ok, not a retailer, I’ll give you that, but hold on) and…wait for it….Shopify!

WHAT! I thought retailers don’t want it? I imagine Andrew saying to this, “we’ll I meant large retailers, like Amazon! I mentioned Amazon!!!

Scot hands Andrew yesterday’s WSJ→

Does this count? Checkmate.♛🏁

Actually, the fallback is always: “That doesn’t count because AGENTIC MUST HANDLE THE ENTIRE BUYING PROCESS!!!!”

On the Amazon point let’s parse through that a bit too. Let’s say Amazon never embraces ChatGPT - I think that’s what they’ll do, until they can’t which will happen faster than everyone thinks.

Like Walmart, Amazon is a mix of 1P and 3P inventory. Some facts:

• 60% of Amazon’s sales are 3P

• Amazon 1P is 12m products -they only stock the top sellers (remember this)

• Amazon 3P is 588m

• 1P is 2% of inventory and 3P is 98%

That inventory is by no means locked in at Amazon - those sellers can sell on Walmart, many do and Walmart can put that inventory on ChatGPT. Also most Amazon sellers have a shopify store, they can put that on ChatGPT directly themselves.

Then remember the 1P product is only 12m SKUs and they are the top sellers. You know who has about 100% coverage of those SKUs? Walmart and Target. When Target joins in, Amazon won’t have any unique products (meaning not also on ChatGPT) with the exception of Amazon’s large and growing private label biz (Amazon Basics, etc.) and Whole Foods. But, those are trade-down things like batteries and commodity goods where unique selection doesn’t matter.

My point is Andrew makes it seem like Amazon’s inventory is unique and defended, it is not. In fact, hmmm, keep reading, I have some more thoughts.

Agentic Naysayers Scoreboard:

• Andrew1: Bad TAM math, weird Agentic definition, doesn’t understand why retailers are accidentally blocking bots today

• Andrew2: He says 0, rest of world says 50-75% of buyers, says retailers like shopify and walmart have zero interest, yet they signed up, word thinking about Agentic Commerce definition.

Score: ACP: 2, Andrews: 0

Overall score: ACP: 4, Naysayers 0 🏆

Seriously though, I hope this review of the Naysayers has helped you think about these arguments and how to approach them when they come up at industry events or inside your company. I’ve now learned to start at the naysayers definitions of things, because it almost always comes back to that. OH, another one I’ve run into is someone thinks when we say Agentic Commerce, we’re talking about the old OpenAI Agent thing that opens a terminal window and does stuff, which is now called Operator. That’s a somewhat bad experiment and 100% not good for buying stuff. My Agentic Commerce always starts in an Answer engine and goes from there. Watch for that one.

As a bonus for long posts, I like to throw a little fun surprise in for those of you that made it this far. Here we go. Deep breath.

If I Were Sam Altman for a Day….

Use your 🪄and make me Sam Altman for the day, aside from REKing Amazon’s RMN and taking out $60B in margin by changing internet traffic patterns to go to their PDPs, the second weak-spot is their biggest strength - the marketplace. You can get 60% of Amazon’s sales and 98% coverage of their SKUs by moving the sellers over to also sell on ChatGPT for cheaper (this is the genius of Marc Lore’s original Jet pitch for you OGs out there 👊).

This is not hard - but you do need a strategic partner. It’s super hard though, because they need to be able to spin up a store quickly and deal with what’s going to be millions of SMBs. Who could we possibly call that knows about that. OH! I’d call Tobe over at Shopify, he can help with this and, what? That’s right, they are already a launch partner of Checkout🤪

ChatGPT and Shopify create a ‘upgrade to ChatGPT’ program where an Amazon seller’s Amazon inventory is pulled over to a Shopify store, that is just a templated home page and a ton of PDPs. The product catalog is then automatically listed on ChatGP.

That gives us the catalog, now how do we make it even more compelling? I know! We offer them a sweetheart deal!! Amazon Sellers have to reduce their prices 5-10% ONLY on their Shopify/ChatGPT store. How can they do this you ask? Well, they can because Amazon charges them 15% AND they all have to spend another 5%+ on RMN over there just to be seen amongst the sea of ad slop. Pay Amazon 20% or pay us 5%, reduce your prices 10% and you net 5% increase in margin - this is like 3X’ing margin for these sellers, they are on-board. Oh yeah, YOU are merchant of record thanks to our friends Stripe (already integrated with Shopify, how handy!) and you get to keep those customers. Amazon doesn’t let you do that, never have never will -remarket to them till you drop friends.

To summarize, what did we just do? We copied 600m SKUs from Amazon to ChatGPT at a 10% discount. We have Amazon-scale selection at lower price and we have distribution. buyers are going to love this. Maybe we run a ‘Save $100 on Prime, come over to ChatGPT Agentic Shopping and also save 10%, price compare to Amazon today!’ TV ads on the super bowl.

Oh wait - what about fulfillment?! Good news friends, you get to keep using FBA, and to encourage you to, we’re going to put a “ChatGPT verified 2-day shipping!” badge on your listings too. Now you have 600m SKUs, lower price, exact same service level and no Prime membership necessary.

That’s Phase I, but I’ll stop there because there’s only one day 😁.

I mention this to help everyone think a little outside the box. We live in a world where the GAM (Google, Amazon, Meta) hasn’t really faced a real competitor. OpenAI is the biggest threat these folks have ever seen and is moving at an unbelievable speed with massive massive resources and if you don’t think Sam Altman and team understand this stuff and see these weakspots, I think you’re going to really be surprised over the next 2-3 years what’s about to happen….

Conclusion to the Trilogy!

Hopefully this series has armed you with all you need to know about ChatGPT Checkout.

• In Part 1 - We started at the top, looked at the consumer experience and the merchant experience. We thought about what it means for merchants and Agentic Commerce. Hopefully I’ve convinced some of you that this is a big deal and then the events of this week solidified that (it’s just the start)

• In Part 2- We got a bit nerdy, talked about specs, data feeds, etc.

• In Part 3 - We covered some of the open questions out there, provided some answers to FAQs we’ve gotten and finally heard from the naysayers to make sure we’re looking at both sides of the argument.

Key Takeaways

• ChatGPT Instant Checkout and ACP is a tectonic shift in the retail and ecommerce World. It’s going to create a huge anti-GAM alliance that now includes ChatGPT, Stripe, Shopify, Walmart and Salesforce with more to come.

• ChatGPT from the CEO down is committed to ChatGPT Checkout being the most merchant friendly solution available for selling online. This is unique and a huge strategic advantage. I believe it could be the key to really giving Amazon some heartache over the next 2-3 years. We’ll see how that unfolds.

• Invest deeply in your product catalog, this is hugely strategic. Do not just dust off your Google Shopping feed and send it to ChatGPT. That is a serious self-inflicted error. Conversely, the biggest strategic move you can make right now is investing in upgrading your product catalog to be 100% Agentic Ready. Before the ChatGPT Checkout news hit, we had just put out Part II of our three-part GenAI Engine Optimization series (part 1 is here and part 2 is here.). In Part III coming next week, we will go into deep details on this. Stay tuned, and read parts 1 and 2 to get ready and understand more of why this matters.

• Don’t believe the naysayers. Their arguments are flawed because they are based on not understanding retail and ecommerce, or very narrow/limited and aggressively applied word/definition thinking. This has been, and will be, proven out by their predictions being incorrect over and over again.

Thanks for reading. If you enjoyed be sure to subscribe and share with a friend or on socials, it’s appreciated!