Retailgentic | Consumer Behavior & Retail Trend

In this episode of the top Agentic Commerce Podcast we explore the powerful shopping powers of the OpenClaw open-source agent. Can the Claw shop on Amazon, shop locally and use the VisionClaw system with META Ray-Ban’s to buy a very hard to search for dog treat? Tune in to find out!

Highlights:
  • What OpenClaw is (and why “persistent” changes everything)
  • The real difference between OpenClaw vs. ChatGPT/Claude as tools
  • How OpenClaw runs (Mac Mini vs. cloud VPS like DigitalOcean)
  • Why “no guardrails” is both the innovation and the risk
  • Risk mitigation: limiting blast radius with single-use virtual cards
  • Why browser-driven commerce is riskier than server-to-server protocols (ACP/UCP)
  • The “skills” layer: how OpenClaw becomes extensible like Lego bricks
  • Why SERP-style APIs beat slow webpage crawling for shopping tasks
  • Live demos: product search, local inventory, and frictionless checkout
  • Using Slack as the agent interface (channels for shopping, health, etc.)
  • The sci-fi moment: Meta Ray-Ban glasses + VisionClaw + Gemini for visual shopping

OpenClaw isn’t the final answer for agentic commerce, security and architecture matter, but it is an incredible preview of where the UX is heading. If you want a glimpse of the “always-on shopping agent” future (and how people are hacking it together right now), this one’s for you. Happy agentic commerce.🦞

Timestamps:

01:38 — OpenClaw takeover begins: why Scot called Ryan as a lifeline
05:26 — What OpenClaw is: persistent agent, open source, “unhinged,” and why that matters
08:21 — The “gigantic loop”: always-on agent behavior (old-school “cron jobs,” new-school autonomy)
09:26 — The killer feature: OpenClaw keeps working after you close your laptop
12:33 — Read-only Gmail as an assistant: catching missed emails + reducing cognitive load
14:35 — The trap: spending time building automation to “save time” (catch-22)
15:33 — Why OpenClaw-style commerce is riskier than ACP/UCP: browser manipulation
18:02 — Gateway Dashboard tour: where OpenClaw is configured (channels, skills, cron jobs)
20:30 — Skill spotlight: Rye for agentic Amazon-friendly purchasing
22:39 — Skill spotlight: SERP API for faster product search + ratings without crawling
24:30 — Skill spotlight: Buy Anything for context-filled checkout (address, identity, card)
28:19 — Demo: “Find me the 5 best-rated 65-inch TVs” (online + local)
35:20 — Live purchase demo: finding USB-C cables, choosing Anker, and the agent checks out
37:05 — “Most frictionless checkout”: no login, agent already knows everything
37:31 — What this means for Alexa: why “available 24/7” changes habits
39:07 — Meta Glasses demo setup: first-person vision commerce 
40:24 — Vision commerce in action: identifies product + searches best price automatically
41:34 — Multi-store results: Amazon + Chewy + Walmart links dropped into Slack

👉 Connect with Ryan: https://www.linkedin.com/in/ryaneade/

🧠 Want to stay ahead in AI commerce? Subscribe and follow along:
📰 Subscribe to the free Substack: retailgentic.ai
📺 Watch episodes on YouTube & Subscribe for updates: https://www.youtube.com/@Retailgentic

Creators and Guests

Host
Scot Wingo
CEO and Co-founder, ReFiBuy.ai
RE
Guest
Ryan Eade

What is Retailgentic | Consumer Behavior & Retail Trend?

Retailgentic, presented by the ReFiBuy Media Network and hosted by Scot Wingo, is where Agentic Commerce meets retail innovation. We help retailers and brands prepare for the future of agent-driven shopping with actionable news, expert insights, dxis, forward-looking predictions, and practical guidance on Agentic Commerce Optimization (ACO).

Speaker 1:

Depending on what you're connecting it to might impact some of that. I still live in the real world. I'm still a real human being. I still very much value my relationships with everyone else. And for me, you know, OpenClaw actually gives me more time to spend with family, with friends, doing stuff I like doing.

Speaker 2:

Welcome to the Retailgentic Podcast, where we explore the intersection of retail, ecommerce, and AI agents, which we call Retailgentic. Our mission is simple. We are shining a bright light on the new agentic AI shopping megatrend which will disrupt the trillion dollar digital retail market. The stakes are high. Google, OpenAI, Microsoft, Perplexity, Amazon, Apple, Meta, and Anthropic are all fighting to be the Consumer AI Shopping Agent winner.

Speaker 2:

There are gonna be some tough decisions for online brands and retailers to make. What happens to checkout? How should payments work? Could you continue to invest in retail media networks? What does the world look like when over half of your traffic is not from humans but from AI agents?

Speaker 2:

These existential questions are the tip of the iceberg. On this podcast, we're talking to industry experts, analysts, and influencers to get answers to these questions, identify trends, and best practices. This is Retailgentic, your podcast partner in navigating our AI shopping agent future. And now here's your host, Wingo.

Speaker 3:

Welcome to this very special edition of Retailgentic. It's an OpenClaw takeover. Everyone get your claws out for this one. In November 2025, Clawdbot launched. Ever since then, every week, I get at least five requests from, avid readers and viewers and listeners that want us to talk about the impact of Clawbot, which is now...

Speaker 3:

It went through a phase where it was called MoltBot, and then it is now called OpenClaw. So I'll... Going forward, call it OpenClaw. I have received a veritable flood of requests for people to, explain how this very exciting open source project could impact Agentic Commerce. It's a pretty technical system.

Speaker 3:

I don't recommend, listeners that are in a corporate environment install it there. This is definitely a home project. It will go into what it is. I don't wanna spend our time here on that, but it has a... It...

Speaker 3:

It's basically doesn't have guardrails, and that gives gives it immense power, but also, you know, a lot of security risk. So be careful with your OpenClaws out there, people. That being said, I did want to demonstrate all the latest cool things that it can do. So I had to call a lifeline from... For someone that I knew had gone pretty far down the OpenClaw rabbit hole and was using it actively.

Speaker 3:

And they were able to, within a couple of hours, do some research and find the very best of the best. And also a really cool demo at the end here that is going to stretch OpenClaw's capabilities and show you a demonstration that looks a little bit like science fiction. So I hope you enjoy this conversation. This is kind of like the bleeding edge area of Agentic Commerce. There is a lot of concern.

Speaker 3:

Our friends at have a really good paper out on this. It talks about how OpenClaw will not be the Agentic Commerce solution. I I agree with that perspective because it does have a lot of security issues. That being said, Ryan actually came up with some really interesting best practices for minimizing those. It doesn't limit them totally, but, and it's not a 100% what you'd want to see out there, but it does minimize it from a consumer's perspective.

Speaker 3:

So this is, if you like this, let us know. We'll be doing more of this type of content. That's like a little bit further out, and, you know, helps us kind of stretch our thinking about where this is going a little bit. I hope you enjoy this one. I had a really great time working with Ryan on it, and I learned a ton.

Speaker 3:

And I think it is really interesting system, I'm gonna be, using my spring break, to the nerdy endeavor. I will be doing is getting my OpenClaw installation going so I can have fun with this new system as well. Enjoy the pod and happy Agentic Commerce. Welcome to this very special OpenClaw edition of Retailgentic. The number one request I have gotten since the OpenClaw craze hit, also known as Clawbot and MoltBot, but now we won't use those anymore because we're 100% OpenClaw, is to have someone come on and talk about how will OpenClaw impact Agentic Commerce.

Speaker 3:

I had to call a lifeline, so called my good friend Ryan Eade. He is a developer extraordinaire and master of all things geeky and nerdy. And I thought if anyone could do this and has the ability to figure this out, it is Ryan Eade. Ryan, welcome to the show.

Speaker 1:

Thank you, Scot. Happy to be here. Happy to share some of the new cool stuff with OpenClaw.

Speaker 3:

Let's talk a little bit about what is OpenClaw. So most people have heard about this in some form, but the simplest way to help you kind of understand it is it is it's an agentic thing. So in our world, an agent is an AI. So it has an LLM based heart always. And then it can go do things.

Speaker 3:

Now OpenClaw is this inspiration by it's an open source project, so he just made it freely available. And what's his name again? Oh, I forget.

Speaker 1:

Peter Steinberger.

Speaker 3:

Peter Steinberger. And he's kind of an entrepreneurial kind of developer type person. He had a a PDF thing that he sold to a big company and has done a lot of other, these kind of, auction real projects. And he basically wanted to have... Today's agents are very tethered.

Speaker 3:

So they have lots of rules and restrictions, and they're constantly pedantically asking you for permission. So one of the things about OpenClaw that I'll say here at the front is it's basically just totally unhinged, meaning like it has no rails and it can go do anything. So because of that, it has a lot of risk, we'll go into some of those in a minute. So this is also why I wanted to have Ryan Eade because I knew he would be able to manage these risks better than me. I would like destroy him.

Speaker 3:

I would like, you know, I would open up my bank account and everything and some bad actor would drain my bank if I did this. So part of, so what you do is you can just go to openclaw.ai and you can download this onto your computer. Don't do that. But that's, it's very, it's relatively easy. You do kind of need some developer skills.

Speaker 3:

What a lot of people do is they actually buy a Mac Mini, which is that little puck looking Mac, and they put it on there. That's actually way too much compute for one of these, and you don't really need that. What we're going to show you here is Ryan went ahead and got a cloud account at one of the... It's like an AWS type company, but it's called Digital Ocean. It's basically $20 a month, which is way cheaper than buying a $4,000 Mac Mini.

Speaker 3:

Some people get really crazy, and they'll get a... You can actually run these on open source models if you don't want to pay for the model. If you do that, you're going need some hardware or a virtual machine that's a little bit more powerful. So a lot of people get a Mac Studio, which is like the $10,000 Mac. The innovation that it has, in addition to being kind of like untethered, meaning it can basically go do anything, is it's extensible through skills.

Speaker 3:

So you can add a bunch of skills to it. It also very cleverly, it has a personality and that is expressed through its soul. So that's kind of interesting. And what's different from it being like something like in your browser or on your computer or like Claude in Excel or an agent like that, that's kind of like attached to something is it's persistent and it lives, you know, either on its own hardware, your hardware or in the cloud, and it's persistent. So that means it's always on.

Speaker 3:

So it's always available. And then what it does is it runs an Agentic loop is what it's called. So it's always there kind of thinking through what to do next. In the old days, we called these cron jobs. I think underneath it may still call it cron at some point.

Speaker 3:

But it's kind of always like kind of thinking about, you know, like, you know, what should I do next? So it's in this kind of like, it's just sitting there looping, waiting for you to give it instructions and do more. And then if you give it work, it will keep working on things for a very, very long time. Some of the risk. So in this demo, does, to make it effective, if you take all the risk out, you've basically neutered it and it's useless.

Speaker 3:

So what Ryan has done is given it access to his email in kind of a read only mode and calendar and these types of things. So he has one that's kind of like his personal assistant. That's not what we're going to demonstrate here. Anything else you want to say about OpenClaw that, you know, to help people orient on what it's been like living with OpenClaw for sixty days?

Speaker 1:

Yeah, yeah. So it's, you've probably read stuff about it online. There's a lot of opinions. A lot of people are saying it's the best new thing in AI. Others are worried that it's dangerous and can cause some security issues for you.

Speaker 1:

I think the big thing that people ask is how is this any different than using Cloud or ChatGPT? Scot really nailed it. It's just that it's persistent and in the cloud. And when I close my computer and walk away, my AI is normally stopped. OpenClaw and your Cloudbot can now ping you later.

Speaker 1:

They can keep working on tasks and ask you questions later. You can be sitting on the couch and get a message back and say, hey, I was working on this, but I need something from you. And so that unlock really kind of turns it beyond a tool that you're using and it's more like a partner. And so when you think about it that way and it's kind of always around doing stuff for you, you can give it a task right before you leave your desk for the day. Hey, I'm leaving work, but work for work for the next few hours on this for me.

Speaker 1:

And it's really been an unlock for me in time savings because I've got kids married lots of lots of stuff going around. I'm always looking for ways to save time, skip some of the busy work. And I think OpenClaw really helps with that. And there are so many tools and skills you can pull down now to connect it. It's like Lego bricks.

Speaker 1:

We always use that metaphor, but it's really anything you can think of. You can probably figure out a way to help it with that task.

Speaker 3:

The people I know that go into the OpenClaw rabbit hole, they go deep and you're kind of, I don't know if the rabbit hole has a bottom, but you're way down there. This happens to everybody because it's kind of life changing in a weird way to like have this persistent assistant that can just do things for you. It's very empowering. Now we're only going to talk about the Agentic Commerce slice of that, but keeping it broad for a minute, how, you know, so so you can interact with it. Once it's all set up, be it on a a local piece of hardware, a Mac mini, a persistent, or or in the cloud, you...

Speaker 3:

And you get it rolling and the Agentic loop is running, then it's available to you twenty four seven. So so you you can be texting it, you can Slack it, you could Microsoft Teams it, you can telegram it, you can signal one of them. Yeah. There's a zillion different ways. You could even talk to it.

Speaker 3:

So a lot of people will connect it. If you have a big commute, they'll set it up in their car and they'll talk to it. Have you, so, so give us some examples of how this has changed your life.

Speaker 1:

I talk a lot about how there's so much great content out today. Like there's, there's just podcasts and articles and sub stacks and, and I can't keep up with all of it. And I don't even have enough time to read all that content. And so one of the things I do is when links come my way or someone texts me something and says, Oh, you've got to read this. I'll throw that to my OpenClaw and get it to queue that up.

Speaker 1:

It'll process those things. And then I have it generate, you know, kind of custom podcasts for me, it'll send me an audio file. And when I'm commuting to the office, I can just pull that up and play. And it's basically generated a little podcast for me that summarizes all the articles I sent it the previous day. That alone has been a huge time savings.

Speaker 1:

And the other thing is it's really good at so I gave it read only access to my personal Gmail. And like all of us, I get too much email. So it's really easy to miss those important ones. But it'll call out and say, Hey, you got this email yesterday, looks like you need to respond to it, but you didn't respond to it. And it kind of reminds me, it's really like in a way it's like having a personal assistant or an EA that is constantly kind of checking around the things that you might have forgotten.

Speaker 1:

And for me, the cognitive load that that's reduced on me constantly trying to keep up with all those things is a game changer, right? We're all so busy. We're always looking at how can I not have to spend time on this? And this has really been helpful for that.

Speaker 3:

But then there's a flip side of this, right? So a lot of people, I read these stories in Silicon Valley where no one shows up at parties anymore because they'd rather hang out with their agent. And there's this like, there becomes this weird psychological thing of, are my agents busy doing things? This particularly impacts developers because they typically will have clogged codes running. And they're building, So developers can now handle five or six projects in parallel.

Speaker 3:

And OpenClaw could be an orchestrator in some of that too. And so now that we're living in this Agendaq world, there's some really interesting behaviors happening. Have you started experiencing any of that, that you have to kind of be like, wait, what am I doing?

Speaker 1:

Yeah. Thankfully, I'm not sitting there waiting on my token to reset for the day. I'm not to the point now where it's like, oh man, I've got three hours and I need to get back in here. But I do know that when you're running through OpenClaw, because it can talk to all the LMs behind the scene, or you can use one of the open models, depending on what you're connecting it to might impact some of that. I still live in the real world.

Speaker 1:

I'm still a real human being. I still very much value my relationships with everyone else. And for me, OpenClaw actually gives me more time to spend with family, with friends, doing stuff I like doing. But yeah, I think the flip side of that, maybe the dirty secret is you get too deep in the rabbit hole. You start feeling like you just need to spend more time with this to get it just a little bit better.

Speaker 1:

And sometimes you can get so obsessed with the time savings. You end up wasting all your free time creating tools to save time. So it's kind of a catch 22 that you spend all this time building automation and now you don't have any free time because you're just sitting there tinkering with it. But I try to balance that, not let it get a hold of me.

Speaker 3:

Okay, good luck. When your wife starts asking if you love you love your OpenClaw more than her, you know you've gone too far.

Speaker 1:

Yeah, she says, I treat her like a human being. I talk to her like a human being. So I think that's part of the flow.

Speaker 3:

Okay, let's talk about Agentic Commerce. Now in these demos, so our friends at Nacuda, which is an online, an Agentic payment company correctly pointed out there's a lot of risk to Agentic Commerce and, OpenClaw. And the reason why is the way OpenClaw interacts with the world is not like what we've seen with ACP and UCP and these server to server protocols that are entirely secure and and dialed in is it's manipulating a browser. And it can manipulate the browser, you know, on the hosted server, or you can connect it to your local browser, and it can do things that way. Connecting it to your local browser is what everyone does because that keeps it from being blocked.

Speaker 3:

When it's on a server, it's much easier for, Amazon is the king of blocking these things. It make... It makes it easy for, you know, to go around that. That's one of the benefits of having it on a local machine like a Mac mini is. It's just kinda looks like a consumer IP.

Speaker 3:

It's almost impossible to block those. So Ryan's gonna show some demos where where he goes and and buy some things, or we're at least going to go shop for some things. Are we going to buy some things?

Speaker 1:

Yeah, we're going to try to buy a product today and we'll see if we get that through.

Speaker 3:

And he's come up with a really clever way of mitigating the risk here. In the corporate and consumer world, you can create virtual cards for a single purpose. So he's using a service that basically allows him to have a single use credit card. And basically... Create a little virtual card, you can program it and say, all right, I want this to have single use.

Speaker 3:

It's got a budget of $50 So if for whatever reason that credit card... It's kind of impossible for it to be hacked. It's still not PCI compliant, so I want to be really clear on that. But it limits your risk to basically zero because, you know, it's single use. So if it did get out for some reason, you would your only downside is $50 not your whole credit.

Speaker 3:

And then also, the chances of it getting out, it's got a chicken and egg problem. So where these things capture is when it's being, you know, communicated across the internet in an unsecured area or, you know, they can kind of intercept that. By the time you do that, you're done buying. It would have to be so, it can't front end that in most cases. Maybe there's some example where it could, but it's limited the risk down to a very, very low amount.

Speaker 3:

So what we're going to do is let's go and just kind of like, let's kind of start in the middle of the onion and then we'll pull out. Like, if you could just log into your Claw and let's kind like see the interface you interact with OpenClaw. So you're going to log into your virtual server and just kind of show us OpenClaw. And then where Ryan spends most of his time interacting with OpenClaw is actually not on this. It's through Slack or a message.

Speaker 3:

So we're going to kind of back out and look at those after this. Alright. I do apologize to our audio listeners. There is going to be a video component to this, but I'll talk you through it. So what we see here...

Speaker 3:

Okay. So what are we looking at, Ryan?

Speaker 1:

So this is what they call the gateway dashboard. So it's part of the OpenClaw platform. So whether you run on a Mac mini or on a server in the cloud, this is kind of like the web based configuration page. So out of the box, this is how you can message with OpenClaw. You don't have to use one of the messaging services out there.

Speaker 1:

This is kind of the built into you. And you can see there's a lot of options. There's different channels and sessions and instances and cron jobs if you want to have it scheduled to ask you questions throughout the day. But then it's got agents and skills and nodes, this is really where you can go and configure all those things. But the funny part of it is since you've got an agent that you're operating with, you really don't use any of those configurations.

Speaker 1:

You're just asking it usually what you want it to do. So you're saying, Hey, install this skill for me, configure this schedule, remind me every morning at 8AM to do this. And it does all that set up for you, but you can go through here and kind of check what all those things are and kind of see what it's building.

Speaker 3:

Yeah. And this is hard to get your head around for a while, if you're not used to this kind of thinking. It can go change itself, right? So, so it, you can, you can text OpenClaw and you can say, I want you to add these six skills and it goes and proves itself. It's kind of like The Matrix, right?

Speaker 3:

Where, where he loads in the skills. And you can actually go tell it to go build things that it then uses. You can go tell it to go build a skill, right? So it is self... It is code that is self coding.

Speaker 3:

It is an agent that can code and improve itself. It is a self improving agent, which it takes a little while. It sounds like science fiction when you say it, but we're living it right now. If, you know, if ten year ago Scot and Ryan had heard this, we wouldn't have believed it. We would have been like, show me that's working.

Speaker 3:

There's like some, there's some, hack in there that you're not really telling me that that's impossible. But this has been made possible through... It all sits on the LLM, but then this kind of framework. By making it unhinged, it allows for it to do more kind of wacky stuff like this. And it's just that's part of the innovation of this that makes it infinitely improvable and interesting.

Speaker 3:

Now let's talk about so staying back on Agentic Commerce. As you researched this, you added some skills. We're So going to talk about some of the skills that you had to add to make it kind of a, give it a little bit more superpower around this one specific task. The first one was called Rye, R Y E. Tell us what Rye does.

Speaker 1:

So Rye is a platform that existed out there that really made it easier to perform a purchase on Amazon through an Agentic flow. And so what that service really does is they're really purchasing something on your behalf. So they're not using your Amazon account, but they allow you to search Amazon, pick an item through a chat flow or a voice flow and then say, yeah, I want to purchase this. And then they ask through the API for things like your shipping information, your contact information, maybe your credit card information which goes through a tokenized flow with Stripe. But it's really aimed at like this interfaced agentic flow because my OpenClaw is talking to it.

Speaker 1:

It knows a lot about me already. It knows my home address. It knows who I am, my name. And because I'm using those kind of one time use cards, it can feed that card in there for me. And so if I say, hey, this is a credit card I want to use for these purchases, which makes that friction of going from what is something that I need?

Speaker 1:

Where can I find it? What's the price to actually buy that for me? Pretty seamless. I will caveat the one cool thing about these tools is you can have it build stuff and create new spills and new integrations on the fly. That's kind of the bleeding edge because of that you run into issues sometimes and sometimes you've got to have it find better paths and maybe something's taking too long.

Speaker 1:

But Rye was a really great service that kind of helped me skip some of that and I was able to search for products and purchase them pretty easily.

Speaker 3:

Yeah, you've inspired me. I have a new life mission. Need to get the Rye CEO on the podcast. So life, new life goal established. I'm a self improving agent.

Speaker 3:

And the other one is so Rye is going to be kind of a it's a very Amazon friendly checkout engine that enables you instead of having to crawl Amazon, and Amazon has all these bot to feeders now, they have somehow figured out how to go around that. It basically lets you abstract that problem and let OpenClaw go and search for you. SERP AI, SERP is a lot of people will know what that is when I say it, but it stands for search engines result page. And why do you need a SERP AI skill?

Speaker 1:

So one of the things that OpenClaw can do out of the box, like you mentioned earlier, it can surf the web. It can crawl pages. If you ask him to search for anything, it's going to use the browser to figure out what you want and it's going to kind of iterate through that web page. It's hard for it to do that. Obviously lots of websites, Amazon included, put all the controls in place to try to keep the AI agents from crawling.

Speaker 1:

And so what SERP API does is it allows it to have a more purpose built API that connects to a lot of different services. So whether that's searching through Amazon or Google Shopping, searching through local stores, Google Places, they've got a ton of integrations built in and it lets your agent and your OpenClaw talk to a single API set that gives it access to a lot more efficient stuff. In my testing, I saw that improved was it was faster at getting ratings back. It could give me the number of ratings. It could give me specific item numbers.

Speaker 1:

It could pull back some product images and it was faster than crawling through all those web pages. Because you can imagine even with an agent, I say, Hey, go find me the best price for this. There's a lot of web pages that's got to open up, iterate through, dig out all that data and pull that back together. And we're in this space where thirty seconds is a long time to wait now, but that does feel like a long time. And SERP API just kind of helped us get to a place where it could do that faster.

Speaker 1:

And so it's free. They do have some limitations around how many times you can use it, but on their free plan, I think it was maybe a 100 searches at a time. It definitely was an improvement.

Speaker 3:

And then last but not least, you have the buy anything skill. What does that do?

Speaker 1:

Yes, so Buy Anything actually ties back into a couple of these other pieces for me and it helps speed up the process of actually checking out without having to do all the what's my address, who am I? And so it ties to some of these other services like Rye and what it's taking is all the context about me that it already knows and filling in those blanks for me without asking me for it. So when I say to purchase something, it can fill in my address, can fill in my credit card number and my temporary one I gave it and kind of makes that a seamless process for the OpenClaw. And because of that, the process of purchasing can be magical. It's kind of sight unseen.

Speaker 1:

You're not navigating to a checkout flow. You're not seeing and confirming everything. So it can feel a little different than what we're used to on the e commerce side. But it is magical when you get through the process and say, well, I just hit purchase and did that. And we're kind of used to some of those now, right?

Speaker 1:

If you've got an Alexa in your house, you can already say, hey, add this to my list or check out with this. But it allows you to leave that ecosystem and still get a lot of that same flow.

Speaker 3:

Those are not your normal glasses. So this is what we call in the biz a teaser. We're gonna have the ultimate... The penultimate demo is Ryan is wearing Ray Ban Meta AI glasses. And we are going to show how you can use those to buy things using visuals.

Speaker 3:

Did you have to install any skills to get that puppy working?

Speaker 1:

Yeah, so there are a lot of integrations you can tie into your Meta Glasses. There's open source product out there called Vision Claw that kind of lets you connect the meta glasses into your OpenClaw and that little connectivity layers is pretty simple. It needs to talk to Gemini to process in the vision. So it's using the Gemini AI from Google to do some real time processing. But once it passes that information over to your OpenClaw, it's like the world is your limit.

Speaker 1:

Whatever you've got integrated, you can start passing that back and forth through. And on the commerce side, there's some cool flows you can do with that. But even tying things into looking at objects in the room and saying, the slide on, turn the slide off. If you've got some smart lights, it really unlocks a lot of cool visual flows. And just having something that's kind of out of sight, out of mind on your face that's got a microphone and speakers built in kind of makes it even more magical.

Speaker 3:

Okay. So let me recap. So you took your basic OpenClaw, you added the Rye skill, Rye. That's a super powerful Amazon checkout agent. You added the SERP AI skill that, that speeds up searches and it's, it's kind of like a cached search experience and, and a more server to server.

Speaker 3:

They They've kind of crawled things and they've got a super fast crawler and it gives your claw very speedy answers. You've got the buy anything that kind of like saves your account information as if you're logged into an Apple or Amazon or Google ecosystem. So you're not constantly having to say you're built to ship to and all that kind of stuff. And then, also, it it helps encode your, your single use credit card. So even...

Speaker 3:

You... You've even reduced your risk even further. And then finally, there's some skills that you had to install, primarily VisionClaw to get the Meta Glasses. So so that's... So so what we're what we're demoing here is basic.

Speaker 3:

You you you could you could recreate this with your basic OpenClaw installed and then these. And in the show notes, I'm gonna put links to all this and kind of give you a... I'm not gonna give you a tutorial because I'm, you're gonna ask questions, I'm not gonna know the answers. I'm talking to the audience. And don't ask Ryan questions.

Speaker 3:

This is not his full time job. And and then then you are good to go from there. So that's that's kinda like the Agentic Commerce superpowers to get to get OpenClaw ready for for what we're gonna do here. Did I miss anything? Absolutely.

Speaker 1:

Nope, I think that's it.

Speaker 3:

All right, so that's a lot of setup. Let's do a demo. So you're in the window here and let's look for something.

Speaker 1:

Yes, maybe look for maybe a 65 inches TV. So this is on the gateway dashboard. So we'll say, help me find a 65 inches TV online or local. We'll say find the five best rated and good pricing. List them back to me.

Speaker 1:

So I'm just typing this in and giving it a command. And this is the most simplistic use case where it will go out and it's going to use some of those skills we've configured to go crawl and figure out what those are. And it's going to use Amazon, it's going to use the SERP API to find some local options and churn through all those different product things and pull back information and then sort that into a list for me. So you could do this today by going to your browser and searching for that. And you might use Google Shopping.

Speaker 1:

You might see a couple other things, but it's kind of doing all that work for me and kind of just pulling this back. And one of the things that is helpful is it already knows kind of where I am. It knows my identity. It's my bot. And so you could see here, scrolled up on me, it pulled in some options online.

Speaker 3:

So it's given us six options and it's giving us very detailed results, right? So it's kind of saying it's giving us reviews, where it's available, the review source, and then a why to buy it type thing.

Speaker 1:

And one of the cool things is even pulled in stock for Best Buy.

Speaker 3:

Yeah. And then at the bottom, it says if you want it locally, here's where to go. Now, it didn't include the pricing. Could you then ask it for the price on that Best Buy item?

Speaker 1:

Yeah, let's see if it can do that since it didn't pull it in. And because agents are, they don't always do things the same way. They kind of do things on what they think is the best pattern. Sometimes you to be more specific with it. And so I'll say, can you pull?

Speaker 1:

And so one of the things I've noticed, I did a couple other shopping flows with this and I appreciated one of the local options it gave me earlier this week was not only did it tell me it was in stock at one of the local Walmarts, which Walmart had it in stock that was on the way to my office from my house. So I didn't tell it to do that. It just knew it knows where my office is and knows where my house is. And it found a store kind of between the two that had it in stock. And I started thinking like, how would I have done that?

Speaker 1:

I could have gone through all those steps myself. I could have said, okay, they've got the best price at Walmart. Now let me go to Walmart. Let me search for which Walmarts are on my commute every day. But I felt like I would never have taken that much time to go through all that flow and it kind of did all that for me.

Speaker 3:

It scrolled, it did this scroll thing again. There it is.

Speaker 1:

So here we go. So we've got best local deal summary. So it is pulling local prices for us now.

Speaker 3:

This is awesome. So if like a hot product dropped, you could just have it checking Best Buy local inventory in real time for you. This is just gonna change everything. You can go have it buy tickets for you.

Speaker 1:

Yeah, I mean, it layers, right? Layers on top of like Best Buy, don't leave your car just to pick up in front of the store with all the stores offering that now. Like I can imagine like hopping in the car from the office, asking it to do this for me, telling it to go ahead and purchase that for local pickup and then picking it up on my way home. It's just such a huge time savings.

Speaker 3:

It'd be a killer for groceries and meal planning too. We have some of that in the core Agentic Commerce world where Chat 50 can talk to Instacart and things, but being unhinged, freeze it to kind of like, you could go use any of your local groceries and fill out the cart and all that kind of stuff.

Speaker 1:

Yeah, looking at your pantry, that's another thing I've been using it for. Just seeing what can I make with what I have and what products are in my pantry right now, which is super cool?

Speaker 3:

Okay. So it has given you... Again, we got... Let's go back up to the Internet availability. So we got local availability.

Speaker 3:

So there's a local angle here that's really cool. And then like, what was the number one on the Internet? Maybe that Samsung on Amazon. Could you like that number three there? Yep.

Speaker 3:

So let's click into that. Could you tell it to go buy that for you? Like, is this where the Rye thing would kick in?

Speaker 1:

So in this view right here, I could ask it to purchase one. I haven't done it through through, like, the OpenClaw control directly. I've done that through Slack.

Speaker 3:

Okay. Let's go let's go out to Slack. So... Okay. Let's...

Speaker 3:

So that was a good example. So that was like... I wanted to kinda like help... Have kinda like start in the middle of the, like, abstraction layer, and then this is like inception. So so that was like the top dream.

Speaker 3:

Now we're gonna go a dream down. So so so you've seen the Ryan was logged in the virtual machine. You've seen it. It's running OpenClaw. You saw the interface there.

Speaker 3:

A lot of people use console, but that's not very user interface friendly. So now we're going to pull back a layer and let's look at one of your favorite ways to communicate with it. You choose.

Speaker 1:

Yeah, so I'm into Slack all day every day. So I use Slack for a lot of work stuff, personal stuff. My kids are in Scouts. We use Slack there too. So Slack is a common tool I'm in.

Speaker 1:

A lot of people with OpenClaw use Telegram. You can also connect it to iMessage, but Slack is just an area that I find myself in a lot. So I thought it was the easiest place for me to interact with it. And because Slack has that idea of channels, you can have dedicated channels for different things. So I've got a health channel that I use to log what I'm eating every day and OpenClaw helps me keep track of that.

Speaker 1:

And then I've got a shopping channel for if I need to purchase stuff or I want to add stuff to my shopping list. And so it lets me kind of have these focused areas. And so these are just some other searches I'd done on like USB C cables. We've got the new iPhones, the lightning cables are out. Our cars have USB C now, so we find ourselves never having enough USB C cables.

Speaker 1:

And so you can kind of do pretty much the same thing we just talked about, like help me find a deal on USB C cables. I'll just say specifically on Amazon, highest rated, we'll say three foot because we're gonna put this one on the car. And it will kind of reach out and do that same flow. So one layer down, the benefit of this is I am on my Mac right now, but Slack is also on my phone. And so a lot of times I just voice chat directly.

Speaker 1:

And so my normal flow for this isn't really typing into it. It's telling it, go search for this for me and using voice to text it processes that command for me. And so that's really helpful for being on the go. My commute's like thirty to forty five minutes depending on traffic. And so that time time I can use to do voice commands.

Speaker 1:

And I can say, add this to my shopping list, add this to this platform. And so that's why I use Slack and why I find it to be helpful. And so it's asking me now, do you want me to order one? And let's just say yes, let's order Anchor. So its top pick was the UGREEN, but I think I like the Anchor brand a little better.

Speaker 1:

So let's see if it can

Speaker 3:

go wrong

Speaker 2:

with this.

Speaker 3:

I'm an anchor enthusiast.

Speaker 1:

It's a brand I trust. I have a lot of stuff from them. And so what it hopefully is going to do, agents are unpredictable, but it's going to use all those skills that I've put together to see if it can do a checkout for me with purchasing this from Amazon. Chain of flow is all happening behind the scenes now.

Speaker 3:

Now in the OpenClaw community, there's a lot of fun that people have. I noticed yours is called Ron Claude Von Don. So so, yeah, so that's fun.

Speaker 1:

Yeah, it's a it's a combination between Ron Swanson and Jean Claude Fandam and a little internal family joke where my wife's southern drawl, my name Ryan sounds like Ron to Siri. So that's how it came up. NC, it fixed something for me. So this is kind of on these kind of cool use cases. So it messed up something, right?

Speaker 1:

Integrations API, if you're trying to do this yourself, you think I've got to code this and I might make a mistake. It had a different field name, we'll say for the Stripe token for my credit card, but it fixed it on its own. It said, Hey, this is wrong, but I know how to fix it. It's going to fix it for me. And then it submitted the order for me and got that placed.

Speaker 1:

And so confirmation email, all the pricing and everything went there together. And so one of the cool things about this is like, I didn't even have to tell it anything. It's already got all that information about me. This is literally the most frictionless checkout flow I've been able to use. No login.

Speaker 1:

It already knows everything.

Speaker 3:

So to recap for those listening, we're in Stripe. Ryan is just chatting with Ron Claw and had a look for some cables that he needed. It splatted out five options very quickly and he said, go buy the Amazon one and it just did it. So, you know, required no no interaction, could have been anywhere. What's this mean for Alexa?

Speaker 3:

Like, do you use your Alexa anymore, or is this kinda replace that functionality?

Speaker 1:

Yeah. So we've got them scattered around the house like many of us do. We still do use Alexa. I would say that for me, like my car has Alexa built in, but I I never found myself using it there as much. And it's not always around.

Speaker 1:

So this is where like the game changer has been for me. If something is available to you some of the time, it's really hard to kind of make that part of your normal habits. But if something is available to you 20 fourseven, no matter where you are, I mean, my phone is always in my pocket. That's kind of what I'm always going to reach for. I've found myself talking to Alexa less and just asking my Ron for it because it's what I'm used to and it works no matter where I am.

Speaker 1:

And So it is kind of changing my personal flow a little bit. I haven't ejected from the Alexa world yet. It's still the quick way, but it is definitely how I'm working with things more.

Speaker 3:

Let's say a brief prayer to the demo gods and let's try showing what's going on with those cool glasses that you're wearing.

Speaker 1:

Alright. So we're gonna get this set up here, make sure our stream is working.

Speaker 3:

So while he's doing that, we've shown how you can use it to look at local inventory and even down to the store level if offered on their website. And we we were looking for a TV, and we narrowed it down to Best Buy, and it was able to show us local availability and pricing at all the stores within an x mile radius of Ryan's house. And then we bought then we pulled back out to Slack, and then we were able to buy things that way. So now we're seeing a livestream. And this is gonna get very meta.

Speaker 3:

It's meta meta, which is also meta. So, you know, we're seeing Ryan's first person view on the screen right now through those glasses that he's wearing. They're... They have a camera in them, a microphone, and then they're talking to him as well. Now will we hear what it says to you, or that's not gonna really come through?

Speaker 1:

We're gonna find out. We're hoping so. We're gonna see very shortly.

Speaker 3:

Okay. Live demo. We're gonna risk it for the biscuit, as my son would say. Oh, look, I can see myself through your eyes. Yeah.

Speaker 1:

It's like back to the inception thing.

Speaker 3:

Yeah. Now I'm now I'm in... Inside the Ryan bot. You're going to go try to buy some products, but you're... Instead of using your voice or your just your voice, you're going to use your eyes to buy a product.

Speaker 3:

So we're going to use augmented reality to buy a product using OpenClaw as engine. Let's try it out. What could go wrong?

Speaker 1:

What could go wrong? What could go wrong? All right. So we're going to go ahead and enable. All right.

Speaker 1:

So I think we're live. So I've got a product here. And so I'm going to ask it if it can search for this product online for me and see if you can find a good price and purchase it for me.

Speaker 3:

Okay. Don't say anything else. I'll talk now so that it doesn't get confused. Sure. I can do that.

Speaker 3:

I see the Verbac Veggie Dent dog chews searching for the best price and purchase options now. Okay. So it's running execute, and we're waiting. This is kind of fun because this is this is actually a hard... I always use these Veggie Dents because it's oddly hard to find them on Amazon because they throw so many ads up.

Speaker 3:

So this is actually kind of a hard one. So we're putting her through her paces here. So the screen sharing froze on that one. Tell us what happened.

Speaker 1:

Well, my side, it searched for it. It pulled up some product links for me on Amazon and is allowing me to purchase them if I want. So I could go ahead and continue through with that. And it gave me, let's see ones it pulled up. It should have also sent it to my Slack.

Speaker 1:

I can search those here. Let me share that. So what I normally do, anytime I do vision search for the glasses, it automatically drops all the links for what it finds right into my Slack. So that if I wanted to click on them and kind of look at them in person, it doesn't have to just read me all the options back. And so these were the options that pulled up.

Speaker 1:

And so different size bags, 11, 22 pounds, 66 pounds. Is that the size of the pet therefore? I was like, that's a lot of veggies. And then the Amazon links and the ratings for them. And so we've got a couple of different options here.

Speaker 1:

Then it also told me it's available in chewy.com and walmart.com. But Amazon items have free delivery by March 4.

Speaker 3:

Let's go ask it in Slack here. How much is it on Chewy? I'm just kind of curious actually. So we want the small dog, which is $27.33 on Amazon. So it's actually clever because, one of the things that throws a lot of agents for a loop is there's three different types.

Speaker 3:

There is fresh Flexins in and the one you held up is fresh. So the fact that it showed all three and then did it by size is actually very, very smart. I wonder if that's the SERP AI thing. That seems like super, that's pretty good. Like, that may be better.

Speaker 3:

It doesn't have both product cards, but that's that's maybe better than kind of core Chatuchitee at Genta Commerce.

Speaker 1:

So there we go. Veggie Dent Fresh, small dog.

Speaker 3:

Interesting. So they're price matching. Sierra Pet at $7.67. Now that's a hallucination right there.

Speaker 1:

Yeah. Odd that it pulls that in.

Speaker 3:

Okay. Well, know you're very busy, so thanks for doing this. So what we demonstrated is a variety of different ways you can use this. You have to have the skills installed, and it's super, super powerful and very good for those everyday essentials. And then if you had something that was hard to find, this could go help you do that.

Speaker 3:

If you had something that was racing the world to kind of go buy the PlayStation six or whatever type thing that, you know, the coolant of the tickle me Elmo craze, where there's always these products that are very hard to find, having this kind of in an agentic loop looking at local store inventory for you and then sending you a text message to go grab it as quick as you could could be a huge advantage. So there's a lot of really interesting capabilities here. And I'm sure because there's this big community and because it's open source, everyone's contributing to this thing and the snowball is getting big really fast. We may have to have you back in, like, three or six months to kinda, like, give us an update on what's going on. But thanks, Ryan Eade.

Speaker 3:

Really appreciate it. We will do a video of that purchase and put it into the podcast so that you can kinda see, you you you know... I think you see that it it worked there, and no wizardry there. It's just kind of a problem with our screen share. Thanks, Ryan.

Speaker 3:

Really appreciate it.

Speaker 1:

Thanks, Scot.