Howie W Ross: I think we've seen incredible development in in the open source models, right? I mean, let's say, I don't know, a year, definitely two two years ago, we we weren't even really considering using, you know, open source, open weight you know, models for, you know, professional coding tasks. You know, we're expect we were experimenting with them, you know, a little bit. but now they're they're they're truly viable.
Caleb Bryant: Well, welcome back to Coffee and Commerce. Of course you've got Caleb and Kayla here and we are joined by Howie Ross of Space Dinosaurs. Howie, so good to see you, my man.
Howie W Ross: Great to see ya.
Caila Schwartz: Nice to meet you, Howie.
Caleb Bryant: And before we get started on why the name Space Dinosaurs, Howie th of course this is all about commerce, but it's also about coffee. So what is in the cup that you've brewed up today?
Howie W Ross: Well, I'm happy to say in addition to being involved in commerce, I'm very into coffee. So daily I you know brew a single origin pour over today. well
Caleb Bryant: Okay.
Howie W Ross: and my favorite is an Ethiopian, because it's super fruity and a lot different than what you'd expect from coffee. But excited to be here to talk about not just coffee but also commerce. How about you guys? You are you guys
Caleb Bryant: I love it. I love it. I am
Howie W Ross: drinking anything today?
Caleb Bryant: I am. Kayla, go ahead, you can share yours.
Caila Schwartz: Okay. I'm not I'm not as as fancy. but my husband did pick up a jug of Duncan's iced coffee.
Caleb Bryant: Ha ha ha
Howie W Ross: Make it up for it in in volume there.
Caila Schwartz: Exactly. You know, sick kids at home, you really need to have heart palpitations to get through it, so
Caleb Bryant: Right. So I'm I've got a new machine. It's this Gagia espresso machine in the back. And so I'm trying out this particular coffee that that I just received, it's by Utopian coffee. it is also Ethiopian, single origin as well. tasting
Howie W Ross: nice. Great.
Caleb Bryant: notes are blackberry, bramble, pop rocks, and birthday cake. It's actually quite delicious.
Howie W Ross: Pop rock. I've never heard that for for a tasting
Caleb Bryant: Yeah.
Howie W Ross: though.
Caleb Bryant: Yeah, it's of course it's I got it in my my Salesforce Blue Yeti Cortado mug. I've had at least three of them so far today. So I'm ready to go.
Howie W Ross: Nice. Yeah. I'm on my second. I'll give a shout
Caila Schwartz: Wow, that's a lot.
Howie W Ross: out to the happy bug where you know I usually get my beans from. but yeah, ready to go.
Caleb Bryant: Did I saw your mug. It says what did you say, We believe or how to believe?
Howie W Ross: I w I I I want to believe it's yeah, reference to the the old the old
Caleb Bryant: I want to believe, there you go. Did you see the latest
Howie W Ross: X Files show.
Caleb Bryant: did you see the the latest Joe Rogan podcast, by the way?
Howie W Ross: I'm not sure if I've seen the quite quite latest, but you know, I see the at least the clips quite frequently.
Caleb Bryant: So the latest one is I just listened to it. I think it was with let's see here. Now I gotta find it. but that's okay. We're we we do these things in real time for a reason. with yeah, David Grush. So the latest the latest Joe Rogan episode is with David Grush, who is ex He's an ex Deep Stater, that was also a a whistleblower on UAP
Howie W Ross: Mm-hmm.
Caleb Bryant: stuff, we'll just say. And if you are if if you are also a believer in in in the unknown, I would I'd definitely encourage you to to take a listen to that episode.
Howie W Ross: man. Yeah, we could take this podcast in a totally different dire on on non commerce related direction. I would be happy to do that. you know, I think there's some really interesting developments there. I mean, I think these days, you know, with the disclosures that have happened, it's it's it's not
Caleb Bryant: Right.
Howie W Ross: a question of, you know, if there are UAPs, we can now firmly I mean, even the government is saying there are things flying around and and and we don't understand them. so you know, really interesting, interesting topic to discuss.
Caleb Bryant: Yeah, yeah, for sure. Which actually
Caila Schwartz: It is kind of crazy.
Caleb Bryant: is a great segue into the name Space Dinosaurs, I think.
Howie W Ross: Okay. So so you know, there's the what as I like to say, there's like there's the story, there's the story, and the story. So we'll start off. so the story goes that there were you know, basically there was a group that was that was brainstorming names. And as, you know, people did years ago, you wanted to con we were trying to come up with a a name that conveyed, you know, that we were going to, you know, take over the world figuratively, of course, but we used to use terms like global. worldwide things like that, right? And so we said, Well what's bigger than that? Space. And as far as we know, there's nothing bigger than space. so we landed on that. and then we were brainstorming, you know, kind of fierce, dominant animals, right? You know, lions, tigers, things like that. And somebody said dinosaurs. And we said, But they're they're extinct. And somebody said, But what if they still exist out in space? And that's where space dinosaurs came from. And and the reception to it has been amazing. People have really reacted to it. Everybody wants to know where the name come from came from.
Caleb Bryant: Well, I love the namespace dinosaurs, and I love it, especially in the context of sort of where we're at in commerce today, because it really does feel like a tale of two things, and that is the evolution of commerce and also the introduction of an entirely new dynamic that you know that w that likely ha you know has been you know, sitting on the sidelines for a while, just sort of in in theory and and also in different forms. but that is that is AI and that is agents, right? Not not that dissimilar from from UAPs. Not that dissimilar from
Howie W Ross: Mm-hmm. Maybe.
Caleb Bryant: you know aliens, agents. You never know. Could be the same thing. We're just
Howie W Ross: It's entirely possible.
Caila Schwartz: I mean, if you look at all the news this week, it's feels like we're, you know, living in the twilight zone, so why not?
Caleb Bryant: Yeah, exactly. Exactly. So but that that brings us to you know a great point, I think, Howie, and and and that is how you know, where we're at in in in this sort of space and time in commerce, and how we as as an ecosystem, both platform, the industry, and and and those that are enablers of technology like like space dinosaurs, how we are how how the world around us is changing and how we are leading that change. Right. I think we, you know, certainly from a from a commerce cloud perspective, we've talked a lot about that on on Coffee and Commerce at you know at length and we'll continue to do so. But really I think that the value that that we have brought forward from a from a platform perspective is enablement to help partners like Space Dinosaurs. To be able to to you know to run fast with really big ideas for commerce. but how has that changed your business, right? Like new platform, new opportunity, new agents, you're bringing your own agents, you're using some of Salesforce agents, the sky's the limit. How you know, how has that changed things for space dinosaurs?
Howie W Ross: Yeah. Yeah, that's a great question. And so there are you know, there's several aspects to this. There so of course, you know, we focus on on e commerce engineering and that has been, you know, dramatically reshaped in the past few years, as I'm sure you know your you and your listeners are aware, via what started with coding assistants and has now progressed to coding agents, and we're now in the multi agent paradigm. where it's not just about coding agents, but we are app actually replicating the entire software development lifecycle using agents. Right. So we s we started it was it was it was just coding. and now you know we've layered in the other aspects of our process. Right. So of course we're now doing code review. We're now doing QA. we're we're you know, we're transcribing because our agents, you know, while they can they can browse the web a little bit, they can consume some video, they still love text. These agents just love text and sometimes we have to kind of make them, you know, consume other formats, whether it be images or video. so, you know, so we're you know transcribing literally, you know, all of our all of our meetings and then, you know, extracting the you know requirements. and direction and feedback, product feedback and you know, website and commerce experience feedback back into those agents from those conversations. So you know, from that perspective, we're now having, you know, really a a a much truer end to end development and engineering, you know, agentic practice rather than just, you know, what started out as as relatively simple code completion. you know, there's numerous other aspects to this, including you know, com things that you know, merchants should be doing and should be aware of. and then of course the impact on the consumer as well. So there's a couple other areas that that we could discuss. but would also love to hear, you know, from you know from your perspective and from from Salesforce's perspective, you know, what is what what do you think is is changing and affecting, you know, your partners?
Caleb Bryant: So I think our goal from a platform standpoint is really to unlock all three and for different reasons. I think merchants are consistently asking us to do that they have to do more with less and they ha but they also have to there is a a massive sort of desire to adopt a genetic technology and but but then do so in ways that are That are aligned to their business KPIs. So they can't just adopt AI for for AI's sake. Having said that, there is this really strong desire to do more, to grow more, to increase margin, but do it with less. And so AI and token consumption is is a big driver of that. Number two, I think for for merchants in the context of deploying new experiences either by way of of their partners like space dinosaurs or their internal merchandisers. there they are placing demands on the market to be able to use trusted agentic technologies to be able to speed up development life cycles, to be able to prompt merchants to to properly merchandise a site, to create new experiences that are that that not only get to market a lot faster but are more flexible in how they are deployed and managed. I think also, you know, building on creating autonomous agents to to be able to continuously monitor and optimize the the site experiences and then in addition to that using agentic technology to also syndicate their brand exposure off site as well. So I think I think, you know, the partnership with Anthropic, our partnership with with OpenAI, continued to grow. and you know, from our perspective, what's really important is that the the merchant has really strong foundation and guardrails. and and architecture, right? The plumbing of commerce to be able to to adapt flexibly to market needs and do so in in sort of near real time. We need to be able to allow for merchants to to be able to do that and but and open themselves up to opportunity and and use the best of what's available to to be able to get it done which is why Not only are we innovating on our own agents like, you know, Carter, our shopping agent, like our merchandising agent, like Agentic Commerce Search, like our agentic order orchestration, and and a whole slew of other stuff that's about to come out. And also, you know, whether it's through merchants or through partners like Space Dinosaurs being able to bring forward other agents like Claude for Commerce agents, right?
Howie W Ross: Mm-hmm.
Caleb Bryant: super cool announcement that we are heavily leaning into both on the platform but then also with our partners and so this whole bring your own agent is is is real right like the platform can consume it and it all it will allow for greater collaboration within a trusted scalable foundation that can can allow merchants to be able to meet the market demands where you know where they need to.
Howie W Ross: Yeah. Yeah, that makes a lot of sense. And I I wanted to jump back and you know comment on on one of the first things you said regarding doing more with less. And you know what we're seeing and what I think the the e-commerce industry and the market at large is is is experiencing as well, which is that well well I do think that's true that you know we are we're we're now expecting much more, right, in terms of velocity, of of not just software. but in you know time to market for for for products, for campaigns, what how we should be able to bring our products live on our on our websites by leveraging AI workflows and so on. And and what we're seeing is that while those expectations have increased, we're actually still seeing significant demand for our work. And I'm you know I'm sure that you and the and and the listeners are aware of you know the stable, if not growing demand for for software engineering and and and truly ai enabled you know workers just workforce you know across the workforce right and so the way that i view it you know there's you know i think you know we're all familiar with that this constant debate about you know is ai going to you know take all of the take all of our jobs take all of the you know the laptop jobs and and so on and the way that I like to frame that is that you know, everyone has basically gotten an instant promotion, right? We are all now managers of our agents, right? So all of our you know, our our individual contributor engineers are now, you know, they're managers of of their agents and their work ha they can their you know productivity has increased dramatically as a result. and we've had to change some of the some of the ways of working. to make everybody, you know, multi-threaded and understand how to manage and balance those workloads better as well, right? Like we used to you know, limit work in progress and scrum and make sure that everybody only had one thing to do and they had to continue to do that thing until it was done. Well now we're giving everybody, you know, three or more things to do at a time because they're gonna be managing agents that are going to be off, you know, working on those tasks and waiting for that that workflow to come back. And I don't think that is a unique to to engineering as well. I think that's across you know the full workforce and certainly people involved in in in in commerce that are working on marketing campaigns or collateral for you know taking product live on site and and so on. So you know personally and and I think I can speak for for my company and you know my team the you know, the productivity increase has been dramatic and it's an amazing time to be, you know, building software and, you know, working on the internet and and working with agents, you know, once again.
Caleb Bryant: Kayla, we gotta get you in
Caila Schwartz: Yeah.
Caleb Bryant: here because I know that you've you've done a ton of research around agentic adoption. what are your thoughts on on the evolution and and the agentic dynamic? Maybe maybe it's an alien dynamic. I don't I don't know. Could be either one. But what do you think?
Caila Schwartz: no, I completely agree.
Caleb Bryant: They could be. We'll just call the Greys. It'll be fine. It's so
Howie W Ross: Yeah, you hear the grays, not the not the insectoids.
Caleb Bryant: Right.
Caila Schwartz: Really go down a rabbit hole. yeah, I completely agree. And and what I'm or a wormhole,
Caleb Bryant: Or a wormhole. Or a wormhole.
Caila Schwartz: yeah, or a black hole, you know.
Howie W Ross: There we
Caleb Bryant: Sorry, go ahead.
Caila Schwartz: just looking at how quickly now people are building e-commerce experiences, how quickly they are able to stand up agents. In fact, just last holiday, right, we saw a couple of brands push out their own shopper agents right before. Cyber Week, which prior to I want to say last year, building and launching anything consumer facing right before you know peak season was unheard of. And so not only is innovation and and and engineering and all of this is getting faster, but we're also stretching the the boundaries of what we originally considered was possible within e-commerce. And what I'm seeing right now, especially on a consumer side, just tremendous adoption, but it's it spans outside of you know what we're seeing within agentic technology, right? I mean this is the role of social and how that is growing and where shoppers are discovering and buying is shifting. And part of it I I believe is because we're seeing this new technology and the consumers getting comfortable with new formats of shopping and that's impacting other channels as well.
Howie W Ross: Yeah. Yeah, I think that's a great point. And, you know, we'd love to to talk a bit more about that changing, you know, consumer and customer behavior and th the change in in product discovery and what I I I think has been, you know, a little bit late to the game, but I think is coming, is that change to the actual purchasing workflow as well. So I think that, you know, many of us and you know consumers in general are getting you know much not not just comfortable, it's just become you know standard if we're going to be considering a large purchase or sometimes even a you know a small a small purchase that we're gonna go agent first or we're gonna go you know AI, LLM, you know, chat first. Whereas previously we would go you know directly to a Google search, directly to to Amazon and and using search, we're much more now so starting that research with with LLMs and and AI chat. which, you know, we can we can touch on, you know, the importance of you know, a we're we're calling it AEO agentic you know engine optimization or geo generative engine optimization. you know, I think that you know w Google is recommending that the you know the same techniques from SEO are are are still applicable, right? That we need to have really high quality semantic markup, as I was, you know, speaking about before. The agents still really like text. They're not necessarily going really look at your web page. So they need to, you know, understand semantically, you know, the various components of that page and and how to navigate it. So we're still seeing good impact and and and lift on occurrences in in AI search and and even sentiment from those type of, you know, what were formerly technical SEO and are now technical geo improvements. Of course, you know, the content component you know, shouldn't be it cannot be understated as well. Right, content is still is still king. We still need to be producing, you know, high quality content for you know about our products, for our brand. And we can do some, you know, some structuring of that content for how LLMs and really how it's because of how users of those LLMs want that content, right? They're asking a question, right? So we should frame, you know, be building that content in that QA, in that question and answer you know, format, at least at least some of it, discovery. you know, we'd love to also talk about, you know, the the developments in in actual product purchasing. but we'd love to hear, you know, from you all about you know if that aligns with what you're seeing and you know what sales how Salesforce is thinking about this.
Caila Schwartz: Yeah, I'm happy to to to chime in on this. having a ton of right now and how we feel. I'm curious what your thoughts are. something that I've heard out in the market is to use human written text on your product detail pages. So don't lean into an AI text for your product descriptions. which is interesting. but I agree what I feel like the challenge for brands is now is that you have to you still have to do SEO Now you have to do AEO. Some of the techniques
Howie W Ross: Yep. Yep.
Caila Schwartz: with AEO are a little bit different than SEO. You now also have to manage your off-brand off-property personas. So how you show up on social media, how you show up on external or third-party review sites and other formats, Reddit. So you have to manage those off property trust signals, which you might have been able to ignore, you know, a year ago or 18 months ago. And then in addition to that, now we have ACP or direct feed integrations. And so you now have to manage that. And you h almost have to do all of it because not only, yes, like ChatGPT, OpenAI will rely more on the feed, but we also have bots and crawlers that are still relying on SEO. When we look at, you know, TikTok and Instagram there and Meta, I'm sorry. they're driving tremendous amounts of traffic.
Howie W Ross: Course. Kayla, I I I I totally agree. you know, and just you know, some some practical tips in this regard, for instance. you know, we use a tool, there's there's several, you know, tools for this, but one that, you know, we've had some success with is called peak peak i, where you know, we will work with customers to determine their priority queries. what they want to come up for in in AI search, which is often different than what we, you know, may think that they want to come up for. and then we analyze that, you know, over time and and can actually, you know, measure over time the you know the the the change. Usually it's you know it's it's positive from the work that we do on the the the technical aspects and then the you know the content generation aspects as well. One one other tip that one of my one of my engineers keyed me in on earlier this week. So I, you know, I'm assume you're all familiar with you know with agent skills and just the concept of you know providing s agents various skills and and personas. One of those skills is called unslop. And so and what it does is, you know, take your text, you know, take the AI generated text and make it seem less AI, more human. Right. So I think that's gonna be, you know, a trend in the in the coming year. Maybe we just shouldn't even be telling people that we're, you know, we're doing that kind of thing. But you know, we like to you know, share these these tips and tricks and best practices when we find them. but yeah, I think that you know, people are increasingly preferring you know, real human genera you know, human content, whether it's a you know a conversation, a podcast like this or a really well written, like personable product description. rather than, you know, a very generic, you know, generated piece of content.
Caleb Bryant: Yeah, totally agree. And and I think that's you know, the opportunity for authenticity is n is never more relevant than it is right now. I think, yes, let's use the best of what's available, but also guess what? The best of what's available is human authenticity too. And that comes out in brands, right? Br brands at its core, they are the ability for a company to to to meet the promise that it has made to the market. And and there is there's a cultural authenticity that is represented in that, you know, that promise. and and and you know, to that point, you know, Howie, you've of course you've worked with a number of brands that on on Commerce Cloud. And y you know, the way that brand certainly there is a a systematic and scalable way that that brands can take advantage of and the types of channels that consumers are are in and how those brands can can deliver on that brand promise within those channels, there is a system and there's a scalable way to do that. And also you know, getting back to the content and the experiences and how brands show up has to you know has to continue. to to represent a a level of authenticity, right? Like nobody wants to hear robots talk to robots. Nobody wants to to see just AI generated images, you know, with within in in terms of how they're experiencing the brand. They they s they want it humans want to talk to humans. And they they also want AI to help with the process and and smooth out the inconveniences, but humans also you know we are naturally designed to to to be able to and to the desire to to connect with other humans or the content that that is representative of of that human authenticity. So as you are like as you are helping brands to you know to use the best of what's available and also represent that brand authenticity, what are some of the ways that you know that you are that you're leveraging some some assets or or general guidance that you're leveraging to help brands to to continue to show up while also, you know, scalably show out.
Howie W Ross: Yeah. Yeah, I think that's a great topic. And and you know, a related topic that I've I've actually spoken about on podcasts before is what I think will be a a r a a rise, an increase in in brand loyalty for the brands that that are truly authentic, right? Because you know, we're you know, in our in our industry we're mostly focused on what happens, you know, after a product is is created, right? On the on the technology and commerce side. But even in you know, on in product development. I think it's never been easier to, you know, to bring, you know, products to market. And I think that, you know, many of them do not necessarily have, you know, the level of quality and authenticity that that people really want. So I think for those brands that are are really, you know, going to stand out and but also, you know, you know, find and really go deep into into a niche, I think there's tremendous opportunity for you know these authentic brands to develop this deep loyalty. so that you know you don't have to, you know, go down the chat GPT rabbit hole to, you know, make a purchasing decision. you can just go back to that that brand that you love and still engage in some of that, you know, that traditional you know browsing and product discovery experience that I I think we still like to do, right? I mean that's why, you know, there's been a change in you know the in in the brick and mortar and the mall ecosystem, but you know, the the the top tier malls, the top tier, you know, online marketplaces, they're still they're still thriving and people are navigating them because we still like that experience of of browsing, finding, and you know, and purchasing something. So, you know, while we're you know, we're encouraging our customers to of course adopt Agenta Commerce and, you know, Kayla, I want you mentioned one of the you know, the the the emerging commerce protocols, which I I think are, you know, a bit nascent at at at this point and we haven't really seen that truly a you know agentic purchasing experience. we we we know it's coming and we're encouraging customers to prepare for it, to invest in it. but also, you know, don't lose sight of, you know, your core commerce shopping experience, your core, you know, customer segment, and considering, you know, that omni channel experience that people are still you know, browsing, discovering, researching online, or flipping that over. And sometimes doing that browsing, doing that discovery in store and then purchasing online. So that you know, that full spectrum, you know, omni-channel experience is still just as relevant as it as it was. And we're going to be soon layering in that, you know, we're just going to say, okay, I saw the thing that I liked in the store, you know, agent, you know, please make that make that purchase for me. But we're not quite there yet. And, you know, we're seeing you know, competition or at least, you know, multiple. protocols, there's MCP, ACP, UCP, from you know the various companies participating. we'll see, you know, which wins and and dominates. I was actually expecting that progress to be faster. I expected that by now that you know we would all be making a lot of purchases through Chat GPT. And I don't I don't think that's quite here yet. And it's you know it's interesting. I'm I'm I'm I'm ready and you know waiting for it. but what about you? What are you all, you know, seeing in in in that regard in terms of you know that the omni channel experience and the agentic purchasing?
Caleb Bryant: think that that customers want to be known across channels and I think that they want to be able to be personalized have personalization represented to them in within those channels in real time but not be subject to you know to a segment I think in addition to that, you know, the the opportunity for both for consumers as well as for merchants is to be able to to to be able to generate some simplicity within the relationship while also to sort of move move the relationship forward in ways that consumers are comfortable with. I think, you know, one example of that, quite frankly, is is the announcement of Muse by Meta. I I think is is a you know that one will will be, in my view, a massive unlock from a from a consumer perspective on on personal agents because you know many consumers have invested a lot in you know, in their personal profile as represented on the meta platform.
Howie W Ross: Mm-hmm.
Caleb Bryant: And and I think, you know, that that one will will continue to open the aperture for you know for agentic trust by consumers as well as how brands interact with you know with with consumers by way of of leveraging you know muse as as the mechanism for doing so. So I yeah, I mean I know Kayla's done, you know, some some digging on this as well. Kayla if you're if you'd like to share some thoughts on that too.
Caila Schwartz: Yeah, I actually have some strong opinions on this.
Howie W Ross: Right.
Caila Schwartz: I I completely agree that there's we're not seeing the desire from consumers to completely offload the shopping journey and automate it. in fact the number one most requested experience when it comes to shopping on platforms like ChatGPT, but any, you know. Agentic search platform is they want to be referred to the brand or retail website to finish the transaction and finalize the research. And I think it
Howie W Ross: Interesting.
Caila Schwartz: speaks to the fact that like entertain shopping is a form of entertainment. there's actually services that I've read about in Asia where people can pretend to buy things online and like you there's they're not actually buying anything, but it They think it feels like they are. They're hitting the buy button, they're transacting, they're watching the the shipment as it pr progresses through tracking to delivery, but there's no actual item. So it's like a exactly there's there's
Howie W Ross: Fantasy shopping. Cool.
Caila Schwartz: an entertainment value to shopping that I think in some ways will never go away. It's just how that entertainment is you know, the format of that entertainment is changing. And I think the first way we see that is with social. Like social is driving a ton of e-commerce traffic right now. We're seeing TikTok and Instagram really grow in terms of e-commerce referred traffic. We're seeing platforms like TikTok shop grow in terms of you know transactions and customer base. so that's one format where we see a shopping evolve. And then in terms of Tools like ChatGPT or Gemini or other third party agents. It's another form of entertainment. In fact, when we look at survey respondents, especially high income earners, they really like the personalization aspect of it. They like to see, you know, what is this agent going to recommend for me based on what it knows about me? and so it it it's it's interesting. So you do have to balance that. And I think Be realistic about the fact that transactions just aren't there yet, but traffic definitely is.
Howie W Ross: Mm-hmm. Yeah, those are great points. And that's, you know, that's really interesting that, you know, first of all, just that you, you know, have and you know have have so much access you know at your fingertips to those to those data sets. And I think that it you know that explains, you know, some of what we're seeing. And I think why maybe there was I believe, you know, some shift in direction on you know, open AI side and you know, potentially some of the other you know, e-commerce platform and you know, and LLM providers. you know, away from that, you know, what we thought was going to be an almost immediate, you know, direct purchasing via LLM. so that's really interesting. And I in and Caleb I I I agree regarding just how just import how important the you know the Muse announcement is not not just to see Facebook getting in the game, but you know, you bring up a really good point about you know, what they know about you, you know, because of the the their advertising business, which has been so successful for years. Not just what you're sharing directly in your profile, but what you're accidentally sharing or inadvertently sharing, you know, based on the ads that, you know, you're you know, you actually watch and and and you respond to and just the sheer, you know, reach and distribution that they have. And you know, I'm sure you've you've you've played around with Muse, I I I have as well. And it seems pretty cool. I kinda like the idea of having an agent that is kind of a little bit more more personal in you know that I'm yeah that I'm not necessarily
Caleb Bryant: Yeah. Yeah.
Howie W Ross: you know using you know you know for for for work. but you know I I you know OpenII has also you know almost immediately fired back with with dots, which is you know s you know a a similar kind of always on in the background in the in in the cloud agent. And I think that's kind of the big development here is that, you know, we're moving away, thank goodness, from these agents, you know, running directly on your, you know, on on your machine to kind of always on and in the cloud and able to do these more long cycle you know, research and potentially, you know, purchasing behaviors there. so, you know, really cool and and interesting developments on that side.
Caleb Bryant: I agree. I mean I think I think what Muse represents, in addition to what I mentioned also is that there will be multiple winners in this market. And there it will not be a duopoly. thankfully there will will will also probably not be regulatory capture in the near future. and
Howie W Ross: We'll see. At least during at least for this there during this administration. Yeah.
Caleb Bryant: At least for now. but I think but I think like, you know, that's a good thing. And and and you know, open source is a great example of that, right? Like you can't put the genie back in the bottle as it relates to open source. Right? There are there are models that are being run on you know on local machines and y you know, and even now, right? Like the just the use of of AI used to be you know, frontier models doing eighty percent of the computing, whereas now open source models are doing eighty roughly eighty percent of the computing and and and growing. And I think but there's a use case for both, right? Like there's a there's a use case for both you know really hard problems to solve where frontier models are are exactly the you know the right thing to do. And then there's also you know significant opportunity for more bespoke you know open source, open weights models. that you know that companies can can design for themselves and and build out contextually relevant skill sets for what they are trying to do. Right. which I know that you guys at at Space Dor Dinosaurs have also experimented with both. So if
Howie W Ross: For sure.
Caleb Bryant: you you know, Howie, however much you wanna share on that I think y you know
Howie W Ross: Yeah.
Caleb Bryant: would be helpful.
Howie W Ross: Yeah, I'd love to. So, you know, so a couple comments on, you know, the, you know, open versus closed source models. I think we've seen incredible development in in the open source models, right? I mean, let's say, I don't know, a year, definitely two two years ago, we we weren't even really considering using, you know, open source, open weight you know, models for, you know, professional coding tasks. You know, we're expect we were experimenting with them, you know, a little bit. but now they're they're they're truly viable. And I believe that if we are not already there, we will in the near future be at the point where frontier models are not necessary for, let's say, ninety n ninety percent of coding tasks, right? Will be able to be done by non-frontier, you know, low-cost models. And it it's an incredible development in the industry. And it's one that, you know, many of us e expected given the way that we saw, you know, the previous developments it just in computing and the internet and and and cloud computing and the rise of of of Linux and open source winning, you know, hands down in that you know, operating system and cloud market. So it makes sense that you know, would go the same direction with open source. And thank goodness because it's, you know, it's driving the cost, it's driving the token cost. The frontier model companies are at least going to have to have offerings that are that are competitive in that market. and I I and Caleb, I think what you're you know, what you're alluding to is, you know, and so we've We have, you know, for for many years, as you can probably tell, have embraced that open source. Our business has been, you know, a tremendous benefactor in the whole industry, not just Linux, but of open source you know, frameworks, you know, React and Next and things of that nature that we've used. So, you know, we're now you know, giving back to the open source community with an agentic coding framework that we've released, which is it's called Dino Stack. And you know, as I was mentioning before, where we've kind of, you know, we've encoded the the software development lifecycle into our agentic engineering practice. We're now you know making that available open source. So it was a framework that you know we developed to use internally because at the time there was nothing there was nothing like it that did you know not just multi agent, but also implemented a workflow, and it's a workflow that is dynamic based on the complexity and the risk of the work. And it includes it does planning, it will do architecture if need be. It of course does coding and will break down tasks into multiple components that can be coded by various agents. then those agents will receive a code review by a by a skeptic, and then that piece of work will also be QA'd. And this is all done in an agenc workflow. to be honest, I thought that By we developed this, it's been out in in open source for a number of months. I thought that by now it would be completely obsolete that other, you know, tools that the the frontier model providers would have released this. And while there are you know similar systems, such as ECC or agent skills, and we have some comparison, you can go to dyno stack. Umai and view the comparison to these other frameworks. there really is nothing quite like it that allows you to just clone this repo. And begin to work. And you can take advantage of that multi-step, multi-agent workflow, and then you know learn the tool and get more sophisticated with it. One of the other really important concepts that it has is this ongoing learning and improvement, right? So at the end of every task, every work session, you tell it, wrap it up and you know, encode your learning, show me what you learned, and then we're gonna write that back into the repo for not just you as the agent. But for all the rest of the team the and the team's other agents to take advantage of. So we'll s we're seeing now, you know, a r a recursive continual improvement loop happening with with agentic engineering as well, which you know, is even more of a benefit to us and of course our our customers.
Caleb Bryant: Why did you think that it was im it was valuable to make that open source to the community?
Howie W Ross: why was it valuable? like I I mean, as I mentioned, we just didn't see anything, you know, quite like it. And we we also didn't necessarily, you know, well, there was a there was
Caleb Bryant: Mm.
Howie W Ross: a potential downside that we felt, you know, we we did have an advantage by by having this system. but also, you know, as I mentioned, we had we had benefited so much and you know the productivity benefits that we had gained, we wanted to share. with the you know, with the larger community and you know, with and and also we we thought that it kinda had a relatively short you know life cycle. So we're like, we gotta put it out now. Let's let people take advantage of it. You know, use it for a short amount of time, use it as inspiration for for what is to come next. But here we are months later and we and and and many others are are continuing to use it and and get value from it.
Caleb Bryant: Great. Amazing. Well, we are almost at an hour. So we should probably wrap it up. But Howie, this is so helpful. And I'd love to have you back on as you continue to see adoption of Dino Stack. I'd love to learn more about just the evolution of of how you know that stack is improving, not just based on adoption of merchants, but also even the the community, I think. We're all in this together, realistically. And you know, the things that we can do to help one another ultimately I think will help you will help our ourselves too. but yeah, it's it's such an exciting time and I'm thankful for folks like yourself, Howie, folks like yourself, Kayla, that are that are doing this thing with me, that we can find a way to to to clear a pathway on behalf of merchants together. It really is an exciting time and I yeah, I can't thank you enough.
Howie W Ross: Well, thank you as well. Thank you so much for for the
Caila Schwartz: Yeah, thank you, Howie.
Howie W Ross: opportunity. It's a it's a real privilege to, you know, get to speak here with you with people that are so well informed on you know what's happening in the market and and and with our customers. And also, you know, just have a passion for you know, for these these customer experiences and AI and technology in general. So it's been a real pleasure. Thank you so much.
Caleb Bryant: Awesome. Thanks, Howard.
Caila Schwartz: Yeah, thank you, Howie. And thank you, Caleb. great talk as always. And I I echo the sentiments that Caleb shared and just say, you know, just to add, one thing we learned from Dreamforce was just how hungry people are for information, for context, for perspective. And so you know As Caleb said, it's an exciting time. It's an exciting time because none of us really know where this is all going to net out. and so to have voices like yours, Howie, to share, you know, just what you're seeing and what you're working on, I think is so incredibly important to the industry. So it was such a pleasure to to get to chat with you. And thank you for thank you both for putting up with my my background noise.
Howie W Ross: No no no problem at all then.
Caleb Bryant: Hey, it it wouldn't be it wouldn't be a coffee and commerce if we didn't have a screaming kid or a dino Dana in the background.
Caila Schwartz: Exactly.
Caleb Bryant: Right?
Howie W Ross: Yeah, it'll be mine next time and I and I hope there is a next time. This has been you know really great. So looking forward to speaking with you all another time.
Caleb Bryant: Amazing. Thank you.
Caleb Bryant: Well, welcome back to Coffee and Commerce. Of course you've got Caleb and Kayla here and we are joined by Howie Ross of Space Dinosaurs. Howie, so good to see you, my man.
Howie W Ross: Great to see ya.
Caila Schwartz: Nice to meet you, Howie.
Caleb Bryant: And before we get started on why the name Space Dinosaurs, Howie th of course this is all about commerce, but it's also about coffee. So what is in the cup that you've brewed up today?
Howie W Ross: Well, I'm happy to say in addition to being involved in commerce, I'm very into coffee. So daily I you know brew a single origin pour over today. well
Caleb Bryant: Okay.
Howie W Ross: and my favorite is an Ethiopian, because it's super fruity and a lot different than what you'd expect from coffee. But excited to be here to talk about not just coffee but also commerce. How about you guys? You are you guys
Caleb Bryant: I love it. I love it. I am
Howie W Ross: drinking anything today?
Caleb Bryant: I am. Kayla, go ahead, you can share yours.
Caila Schwartz: Okay. I'm not I'm not as as fancy. but my husband did pick up a jug of Duncan's iced coffee.
Caleb Bryant: Ha ha ha
Howie W Ross: Make it up for it in in volume there.
Caila Schwartz: Exactly. You know, sick kids at home, you really need to have heart palpitations to get through it, so
Caleb Bryant: Right. So I'm I've got a new machine. It's this Gagia espresso machine in the back. And so I'm trying out this particular coffee that that I just received, it's by Utopian coffee. it is also Ethiopian, single origin as well. tasting
Howie W Ross: nice. Great.
Caleb Bryant: notes are blackberry, bramble, pop rocks, and birthday cake. It's actually quite delicious.
Howie W Ross: Pop rock. I've never heard that for for a tasting
Caleb Bryant: Yeah.
Howie W Ross: though.
Caleb Bryant: Yeah, it's of course it's I got it in my my Salesforce Blue Yeti Cortado mug. I've had at least three of them so far today. So I'm ready to go.
Howie W Ross: Nice. Yeah. I'm on my second. I'll give a shout
Caila Schwartz: Wow, that's a lot.
Howie W Ross: out to the happy bug where you know I usually get my beans from. but yeah, ready to go.
Caleb Bryant: Did I saw your mug. It says what did you say, We believe or how to believe?
Howie W Ross: I w I I I want to believe it's yeah, reference to the the old the old
Caleb Bryant: I want to believe, there you go. Did you see the latest
Howie W Ross: X Files show.
Caleb Bryant: did you see the the latest Joe Rogan podcast, by the way?
Howie W Ross: I'm not sure if I've seen the quite quite latest, but you know, I see the at least the clips quite frequently.
Caleb Bryant: So the latest one is I just listened to it. I think it was with let's see here. Now I gotta find it. but that's okay. We're we we do these things in real time for a reason. with yeah, David Grush. So the latest the latest Joe Rogan episode is with David Grush, who is ex He's an ex Deep Stater, that was also a a whistleblower on UAP
Howie W Ross: Mm-hmm.
Caleb Bryant: stuff, we'll just say. And if you are if if you are also a believer in in in the unknown, I would I'd definitely encourage you to to take a listen to that episode.
Howie W Ross: man. Yeah, we could take this podcast in a totally different dire on on non commerce related direction. I would be happy to do that. you know, I think there's some really interesting developments there. I mean, I think these days, you know, with the disclosures that have happened, it's it's it's not
Caleb Bryant: Right.
Howie W Ross: a question of, you know, if there are UAPs, we can now firmly I mean, even the government is saying there are things flying around and and and we don't understand them. so you know, really interesting, interesting topic to discuss.
Caleb Bryant: Yeah, yeah, for sure. Which actually
Caila Schwartz: It is kind of crazy.
Caleb Bryant: is a great segue into the name Space Dinosaurs, I think.
Howie W Ross: Okay. So so you know, there's the what as I like to say, there's like there's the story, there's the story, and the story. So we'll start off. so the story goes that there were you know, basically there was a group that was that was brainstorming names. And as, you know, people did years ago, you wanted to con we were trying to come up with a a name that conveyed, you know, that we were going to, you know, take over the world figuratively, of course, but we used to use terms like global. worldwide things like that, right? And so we said, Well what's bigger than that? Space. And as far as we know, there's nothing bigger than space. so we landed on that. and then we were brainstorming, you know, kind of fierce, dominant animals, right? You know, lions, tigers, things like that. And somebody said dinosaurs. And we said, But they're they're extinct. And somebody said, But what if they still exist out in space? And that's where space dinosaurs came from. And and the reception to it has been amazing. People have really reacted to it. Everybody wants to know where the name come from came from.
Caleb Bryant: Well, I love the namespace dinosaurs, and I love it, especially in the context of sort of where we're at in commerce today, because it really does feel like a tale of two things, and that is the evolution of commerce and also the introduction of an entirely new dynamic that you know that w that likely ha you know has been you know, sitting on the sidelines for a while, just sort of in in theory and and also in different forms. but that is that is AI and that is agents, right? Not not that dissimilar from from UAPs. Not that dissimilar from
Howie W Ross: Mm-hmm. Maybe.
Caleb Bryant: you know aliens, agents. You never know. Could be the same thing. We're just
Howie W Ross: It's entirely possible.
Caila Schwartz: I mean, if you look at all the news this week, it's feels like we're, you know, living in the twilight zone, so why not?
Caleb Bryant: Yeah, exactly. Exactly. So but that that brings us to you know a great point, I think, Howie, and and and that is how you know, where we're at in in in this sort of space and time in commerce, and how we as as an ecosystem, both platform, the industry, and and and those that are enablers of technology like like space dinosaurs, how we are how how the world around us is changing and how we are leading that change. Right. I think we, you know, certainly from a from a commerce cloud perspective, we've talked a lot about that on on Coffee and Commerce at you know at length and we'll continue to do so. But really I think that the value that that we have brought forward from a from a platform perspective is enablement to help partners like Space Dinosaurs. To be able to to you know to run fast with really big ideas for commerce. but how has that changed your business, right? Like new platform, new opportunity, new agents, you're bringing your own agents, you're using some of Salesforce agents, the sky's the limit. How you know, how has that changed things for space dinosaurs?
Howie W Ross: Yeah. Yeah, that's a great question. And so there are you know, there's several aspects to this. There so of course, you know, we focus on on e commerce engineering and that has been, you know, dramatically reshaped in the past few years, as I'm sure you know your you and your listeners are aware, via what started with coding assistants and has now progressed to coding agents, and we're now in the multi agent paradigm. where it's not just about coding agents, but we are app actually replicating the entire software development lifecycle using agents. Right. So we s we started it was it was it was just coding. and now you know we've layered in the other aspects of our process. Right. So of course we're now doing code review. We're now doing QA. we're we're you know, we're transcribing because our agents, you know, while they can they can browse the web a little bit, they can consume some video, they still love text. These agents just love text and sometimes we have to kind of make them, you know, consume other formats, whether it be images or video. so, you know, so we're you know transcribing literally, you know, all of our all of our meetings and then, you know, extracting the you know requirements. and direction and feedback, product feedback and you know, website and commerce experience feedback back into those agents from those conversations. So you know, from that perspective, we're now having, you know, really a a a much truer end to end development and engineering, you know, agentic practice rather than just, you know, what started out as as relatively simple code completion. you know, there's numerous other aspects to this, including you know, com things that you know, merchants should be doing and should be aware of. and then of course the impact on the consumer as well. So there's a couple other areas that that we could discuss. but would also love to hear, you know, from you know from your perspective and from from Salesforce's perspective, you know, what is what what do you think is is changing and affecting, you know, your partners?
Caleb Bryant: So I think our goal from a platform standpoint is really to unlock all three and for different reasons. I think merchants are consistently asking us to do that they have to do more with less and they ha but they also have to there is a a massive sort of desire to adopt a genetic technology and but but then do so in ways that are That are aligned to their business KPIs. So they can't just adopt AI for for AI's sake. Having said that, there is this really strong desire to do more, to grow more, to increase margin, but do it with less. And so AI and token consumption is is a big driver of that. Number two, I think for for merchants in the context of deploying new experiences either by way of of their partners like space dinosaurs or their internal merchandisers. there they are placing demands on the market to be able to use trusted agentic technologies to be able to speed up development life cycles, to be able to prompt merchants to to properly merchandise a site, to create new experiences that are that that not only get to market a lot faster but are more flexible in how they are deployed and managed. I think also, you know, building on creating autonomous agents to to be able to continuously monitor and optimize the the site experiences and then in addition to that using agentic technology to also syndicate their brand exposure off site as well. So I think I think, you know, the partnership with Anthropic, our partnership with with OpenAI, continued to grow. and you know, from our perspective, what's really important is that the the merchant has really strong foundation and guardrails. and and architecture, right? The plumbing of commerce to be able to to adapt flexibly to market needs and do so in in sort of near real time. We need to be able to allow for merchants to to be able to do that and but and open themselves up to opportunity and and use the best of what's available to to be able to get it done which is why Not only are we innovating on our own agents like, you know, Carter, our shopping agent, like our merchandising agent, like Agentic Commerce Search, like our agentic order orchestration, and and a whole slew of other stuff that's about to come out. And also, you know, whether it's through merchants or through partners like Space Dinosaurs being able to bring forward other agents like Claude for Commerce agents, right?
Howie W Ross: Mm-hmm.
Caleb Bryant: super cool announcement that we are heavily leaning into both on the platform but then also with our partners and so this whole bring your own agent is is is real right like the platform can consume it and it all it will allow for greater collaboration within a trusted scalable foundation that can can allow merchants to be able to meet the market demands where you know where they need to.
Howie W Ross: Yeah. Yeah, that makes a lot of sense. And I I wanted to jump back and you know comment on on one of the first things you said regarding doing more with less. And you know what we're seeing and what I think the the e-commerce industry and the market at large is is is experiencing as well, which is that well well I do think that's true that you know we are we're we're now expecting much more, right, in terms of velocity, of of not just software. but in you know time to market for for for products, for campaigns, what how we should be able to bring our products live on our on our websites by leveraging AI workflows and so on. And and what we're seeing is that while those expectations have increased, we're actually still seeing significant demand for our work. And I'm you know I'm sure that you and the and and the listeners are aware of you know the stable, if not growing demand for for software engineering and and and truly ai enabled you know workers just workforce you know across the workforce right and so the way that i view it you know there's you know i think you know we're all familiar with that this constant debate about you know is ai going to you know take all of the take all of our jobs take all of the you know the laptop jobs and and so on and the way that I like to frame that is that you know, everyone has basically gotten an instant promotion, right? We are all now managers of our agents, right? So all of our you know, our our individual contributor engineers are now, you know, they're managers of of their agents and their work ha they can their you know productivity has increased dramatically as a result. and we've had to change some of the some of the ways of working. to make everybody, you know, multi-threaded and understand how to manage and balance those workloads better as well, right? Like we used to you know, limit work in progress and scrum and make sure that everybody only had one thing to do and they had to continue to do that thing until it was done. Well now we're giving everybody, you know, three or more things to do at a time because they're gonna be managing agents that are going to be off, you know, working on those tasks and waiting for that that workflow to come back. And I don't think that is a unique to to engineering as well. I think that's across you know the full workforce and certainly people involved in in in in commerce that are working on marketing campaigns or collateral for you know taking product live on site and and so on. So you know personally and and I think I can speak for for my company and you know my team the you know, the productivity increase has been dramatic and it's an amazing time to be, you know, building software and, you know, working on the internet and and working with agents, you know, once again.
Caleb Bryant: Kayla, we gotta get you in
Caila Schwartz: Yeah.
Caleb Bryant: here because I know that you've you've done a ton of research around agentic adoption. what are your thoughts on on the evolution and and the agentic dynamic? Maybe maybe it's an alien dynamic. I don't I don't know. Could be either one. But what do you think?
Caila Schwartz: no, I completely agree.
Caleb Bryant: They could be. We'll just call the Greys. It'll be fine. It's so
Howie W Ross: Yeah, you hear the grays, not the not the insectoids.
Caleb Bryant: Right.
Caila Schwartz: Really go down a rabbit hole. yeah, I completely agree. And and what I'm or a wormhole,
Caleb Bryant: Or a wormhole. Or a wormhole.
Caila Schwartz: yeah, or a black hole, you know.
Howie W Ross: There we
Caleb Bryant: Sorry, go ahead.
Caila Schwartz: just looking at how quickly now people are building e-commerce experiences, how quickly they are able to stand up agents. In fact, just last holiday, right, we saw a couple of brands push out their own shopper agents right before. Cyber Week, which prior to I want to say last year, building and launching anything consumer facing right before you know peak season was unheard of. And so not only is innovation and and and engineering and all of this is getting faster, but we're also stretching the the boundaries of what we originally considered was possible within e-commerce. And what I'm seeing right now, especially on a consumer side, just tremendous adoption, but it's it spans outside of you know what we're seeing within agentic technology, right? I mean this is the role of social and how that is growing and where shoppers are discovering and buying is shifting. And part of it I I believe is because we're seeing this new technology and the consumers getting comfortable with new formats of shopping and that's impacting other channels as well.
Howie W Ross: Yeah. Yeah, I think that's a great point. And, you know, we'd love to to talk a bit more about that changing, you know, consumer and customer behavior and th the change in in product discovery and what I I I think has been, you know, a little bit late to the game, but I think is coming, is that change to the actual purchasing workflow as well. So I think that, you know, many of us and you know consumers in general are getting you know much not not just comfortable, it's just become you know standard if we're going to be considering a large purchase or sometimes even a you know a small a small purchase that we're gonna go agent first or we're gonna go you know AI, LLM, you know, chat first. Whereas previously we would go you know directly to a Google search, directly to to Amazon and and using search, we're much more now so starting that research with with LLMs and and AI chat. which, you know, we can we can touch on, you know, the importance of you know, a we're we're calling it AEO agentic you know engine optimization or geo generative engine optimization. you know, I think that you know w Google is recommending that the you know the same techniques from SEO are are are still applicable, right? That we need to have really high quality semantic markup, as I was, you know, speaking about before. The agents still really like text. They're not necessarily going really look at your web page. So they need to, you know, understand semantically, you know, the various components of that page and and how to navigate it. So we're still seeing good impact and and and lift on occurrences in in AI search and and even sentiment from those type of, you know, what were formerly technical SEO and are now technical geo improvements. Of course, you know, the content component you know, shouldn't be it cannot be understated as well. Right, content is still is still king. We still need to be producing, you know, high quality content for you know about our products, for our brand. And we can do some, you know, some structuring of that content for how LLMs and really how it's because of how users of those LLMs want that content, right? They're asking a question, right? So we should frame, you know, be building that content in that QA, in that question and answer you know, format, at least at least some of it, discovery. you know, we'd love to also talk about, you know, the the developments in in actual product purchasing. but we'd love to hear, you know, from you all about you know if that aligns with what you're seeing and you know what sales how Salesforce is thinking about this.
Caila Schwartz: Yeah, I'm happy to to to chime in on this. having a ton of right now and how we feel. I'm curious what your thoughts are. something that I've heard out in the market is to use human written text on your product detail pages. So don't lean into an AI text for your product descriptions. which is interesting. but I agree what I feel like the challenge for brands is now is that you have to you still have to do SEO Now you have to do AEO. Some of the techniques
Howie W Ross: Yep. Yep.
Caila Schwartz: with AEO are a little bit different than SEO. You now also have to manage your off-brand off-property personas. So how you show up on social media, how you show up on external or third-party review sites and other formats, Reddit. So you have to manage those off property trust signals, which you might have been able to ignore, you know, a year ago or 18 months ago. And then in addition to that, now we have ACP or direct feed integrations. And so you now have to manage that. And you h almost have to do all of it because not only, yes, like ChatGPT, OpenAI will rely more on the feed, but we also have bots and crawlers that are still relying on SEO. When we look at, you know, TikTok and Instagram there and Meta, I'm sorry. they're driving tremendous amounts of traffic.
Howie W Ross: Course. Kayla, I I I I totally agree. you know, and just you know, some some practical tips in this regard, for instance. you know, we use a tool, there's there's several, you know, tools for this, but one that, you know, we've had some success with is called peak peak i, where you know, we will work with customers to determine their priority queries. what they want to come up for in in AI search, which is often different than what we, you know, may think that they want to come up for. and then we analyze that, you know, over time and and can actually, you know, measure over time the you know the the the change. Usually it's you know it's it's positive from the work that we do on the the the technical aspects and then the you know the content generation aspects as well. One one other tip that one of my one of my engineers keyed me in on earlier this week. So I, you know, I'm assume you're all familiar with you know with agent skills and just the concept of you know providing s agents various skills and and personas. One of those skills is called unslop. And so and what it does is, you know, take your text, you know, take the AI generated text and make it seem less AI, more human. Right. So I think that's gonna be, you know, a trend in the in the coming year. Maybe we just shouldn't even be telling people that we're, you know, we're doing that kind of thing. But you know, we like to you know, share these these tips and tricks and best practices when we find them. but yeah, I think that you know, people are increasingly preferring you know, real human genera you know, human content, whether it's a you know a conversation, a podcast like this or a really well written, like personable product description. rather than, you know, a very generic, you know, generated piece of content.
Caleb Bryant: Yeah, totally agree. And and I think that's you know, the opportunity for authenticity is n is never more relevant than it is right now. I think, yes, let's use the best of what's available, but also guess what? The best of what's available is human authenticity too. And that comes out in brands, right? Br brands at its core, they are the ability for a company to to to meet the promise that it has made to the market. And and there is there's a cultural authenticity that is represented in that, you know, that promise. and and and you know, to that point, you know, Howie, you've of course you've worked with a number of brands that on on Commerce Cloud. And y you know, the way that brand certainly there is a a systematic and scalable way that that brands can take advantage of and the types of channels that consumers are are in and how those brands can can deliver on that brand promise within those channels, there is a system and there's a scalable way to do that. And also you know, getting back to the content and the experiences and how brands show up has to you know has to continue. to to represent a a level of authenticity, right? Like nobody wants to hear robots talk to robots. Nobody wants to to see just AI generated images, you know, with within in in terms of how they're experiencing the brand. They they s they want it humans want to talk to humans. And they they also want AI to help with the process and and smooth out the inconveniences, but humans also you know we are naturally designed to to to be able to and to the desire to to connect with other humans or the content that that is representative of of that human authenticity. So as you are like as you are helping brands to you know to use the best of what's available and also represent that brand authenticity, what are some of the ways that you know that you are that you're leveraging some some assets or or general guidance that you're leveraging to help brands to to continue to show up while also, you know, scalably show out.
Howie W Ross: Yeah. Yeah, I think that's a great topic. And and you know, a related topic that I've I've actually spoken about on podcasts before is what I think will be a a r a a rise, an increase in in brand loyalty for the brands that that are truly authentic, right? Because you know, we're you know, in our in our industry we're mostly focused on what happens, you know, after a product is is created, right? On the on the technology and commerce side. But even in you know, on in product development. I think it's never been easier to, you know, to bring, you know, products to market. And I think that, you know, many of them do not necessarily have, you know, the level of quality and authenticity that that people really want. So I think for those brands that are are really, you know, going to stand out and but also, you know, you know, find and really go deep into into a niche, I think there's tremendous opportunity for you know these authentic brands to develop this deep loyalty. so that you know you don't have to, you know, go down the chat GPT rabbit hole to, you know, make a purchasing decision. you can just go back to that that brand that you love and still engage in some of that, you know, that traditional you know browsing and product discovery experience that I I think we still like to do, right? I mean that's why, you know, there's been a change in you know the in in the brick and mortar and the mall ecosystem, but you know, the the the top tier malls, the top tier, you know, online marketplaces, they're still they're still thriving and people are navigating them because we still like that experience of of browsing, finding, and you know, and purchasing something. So, you know, while we're you know, we're encouraging our customers to of course adopt Agenta Commerce and, you know, Kayla, I want you mentioned one of the you know, the the the emerging commerce protocols, which I I think are, you know, a bit nascent at at at this point and we haven't really seen that truly a you know agentic purchasing experience. we we we know it's coming and we're encouraging customers to prepare for it, to invest in it. but also, you know, don't lose sight of, you know, your core commerce shopping experience, your core, you know, customer segment, and considering, you know, that omni channel experience that people are still you know, browsing, discovering, researching online, or flipping that over. And sometimes doing that browsing, doing that discovery in store and then purchasing online. So that you know, that full spectrum, you know, omni-channel experience is still just as relevant as it as it was. And we're going to be soon layering in that, you know, we're just going to say, okay, I saw the thing that I liked in the store, you know, agent, you know, please make that make that purchase for me. But we're not quite there yet. And, you know, we're seeing you know, competition or at least, you know, multiple. protocols, there's MCP, ACP, UCP, from you know the various companies participating. we'll see, you know, which wins and and dominates. I was actually expecting that progress to be faster. I expected that by now that you know we would all be making a lot of purchases through Chat GPT. And I don't I don't think that's quite here yet. And it's you know it's interesting. I'm I'm I'm I'm ready and you know waiting for it. but what about you? What are you all, you know, seeing in in in that regard in terms of you know that the omni channel experience and the agentic purchasing?
Caleb Bryant: think that that customers want to be known across channels and I think that they want to be able to be personalized have personalization represented to them in within those channels in real time but not be subject to you know to a segment I think in addition to that, you know, the the opportunity for both for consumers as well as for merchants is to be able to to to be able to generate some simplicity within the relationship while also to sort of move move the relationship forward in ways that consumers are comfortable with. I think, you know, one example of that, quite frankly, is is the announcement of Muse by Meta. I I think is is a you know that one will will be, in my view, a massive unlock from a from a consumer perspective on on personal agents because you know many consumers have invested a lot in you know, in their personal profile as represented on the meta platform.
Howie W Ross: Mm-hmm.
Caleb Bryant: And and I think, you know, that that one will will continue to open the aperture for you know for agentic trust by consumers as well as how brands interact with you know with with consumers by way of of leveraging you know muse as as the mechanism for doing so. So I yeah, I mean I know Kayla's done, you know, some some digging on this as well. Kayla if you're if you'd like to share some thoughts on that too.
Caila Schwartz: Yeah, I actually have some strong opinions on this.
Howie W Ross: Right.
Caila Schwartz: I I completely agree that there's we're not seeing the desire from consumers to completely offload the shopping journey and automate it. in fact the number one most requested experience when it comes to shopping on platforms like ChatGPT, but any, you know. Agentic search platform is they want to be referred to the brand or retail website to finish the transaction and finalize the research. And I think it
Howie W Ross: Interesting.
Caila Schwartz: speaks to the fact that like entertain shopping is a form of entertainment. there's actually services that I've read about in Asia where people can pretend to buy things online and like you there's they're not actually buying anything, but it They think it feels like they are. They're hitting the buy button, they're transacting, they're watching the the shipment as it pr progresses through tracking to delivery, but there's no actual item. So it's like a exactly there's there's
Howie W Ross: Fantasy shopping. Cool.
Caila Schwartz: an entertainment value to shopping that I think in some ways will never go away. It's just how that entertainment is you know, the format of that entertainment is changing. And I think the first way we see that is with social. Like social is driving a ton of e-commerce traffic right now. We're seeing TikTok and Instagram really grow in terms of e-commerce referred traffic. We're seeing platforms like TikTok shop grow in terms of you know transactions and customer base. so that's one format where we see a shopping evolve. And then in terms of Tools like ChatGPT or Gemini or other third party agents. It's another form of entertainment. In fact, when we look at survey respondents, especially high income earners, they really like the personalization aspect of it. They like to see, you know, what is this agent going to recommend for me based on what it knows about me? and so it it it's it's interesting. So you do have to balance that. And I think Be realistic about the fact that transactions just aren't there yet, but traffic definitely is.
Howie W Ross: Mm-hmm. Yeah, those are great points. And that's, you know, that's really interesting that, you know, first of all, just that you, you know, have and you know have have so much access you know at your fingertips to those to those data sets. And I think that it you know that explains, you know, some of what we're seeing. And I think why maybe there was I believe, you know, some shift in direction on you know, open AI side and you know, potentially some of the other you know, e-commerce platform and you know, and LLM providers. you know, away from that, you know, what we thought was going to be an almost immediate, you know, direct purchasing via LLM. so that's really interesting. And I in and Caleb I I I agree regarding just how just import how important the you know the Muse announcement is not not just to see Facebook getting in the game, but you know, you bring up a really good point about you know, what they know about you, you know, because of the the their advertising business, which has been so successful for years. Not just what you're sharing directly in your profile, but what you're accidentally sharing or inadvertently sharing, you know, based on the ads that, you know, you're you know, you actually watch and and and you respond to and just the sheer, you know, reach and distribution that they have. And you know, I'm sure you've you've you've played around with Muse, I I I have as well. And it seems pretty cool. I kinda like the idea of having an agent that is kind of a little bit more more personal in you know that I'm yeah that I'm not necessarily
Caleb Bryant: Yeah. Yeah.
Howie W Ross: you know using you know you know for for for work. but you know I I you know OpenII has also you know almost immediately fired back with with dots, which is you know s you know a a similar kind of always on in the background in the in in the cloud agent. And I think that's kind of the big development here is that, you know, we're moving away, thank goodness, from these agents, you know, running directly on your, you know, on on your machine to kind of always on and in the cloud and able to do these more long cycle you know, research and potentially, you know, purchasing behaviors there. so, you know, really cool and and interesting developments on that side.
Caleb Bryant: I agree. I mean I think I think what Muse represents, in addition to what I mentioned also is that there will be multiple winners in this market. And there it will not be a duopoly. thankfully there will will will also probably not be regulatory capture in the near future. and
Howie W Ross: We'll see. At least during at least for this there during this administration. Yeah.
Caleb Bryant: At least for now. but I think but I think like, you know, that's a good thing. And and and you know, open source is a great example of that, right? Like you can't put the genie back in the bottle as it relates to open source. Right? There are there are models that are being run on you know on local machines and y you know, and even now, right? Like the just the use of of AI used to be you know, frontier models doing eighty percent of the computing, whereas now open source models are doing eighty roughly eighty percent of the computing and and and growing. And I think but there's a use case for both, right? Like there's a there's a use case for both you know really hard problems to solve where frontier models are are exactly the you know the right thing to do. And then there's also you know significant opportunity for more bespoke you know open source, open weights models. that you know that companies can can design for themselves and and build out contextually relevant skill sets for what they are trying to do. Right. which I know that you guys at at Space Dor Dinosaurs have also experimented with both. So if
Howie W Ross: For sure.
Caleb Bryant: you you know, Howie, however much you wanna share on that I think y you know
Howie W Ross: Yeah.
Caleb Bryant: would be helpful.
Howie W Ross: Yeah, I'd love to. So, you know, so a couple comments on, you know, the, you know, open versus closed source models. I think we've seen incredible development in in the open source models, right? I mean, let's say, I don't know, a year, definitely two two years ago, we we weren't even really considering using, you know, open source, open weight you know, models for, you know, professional coding tasks. You know, we're expect we were experimenting with them, you know, a little bit. but now they're they're they're truly viable. And I believe that if we are not already there, we will in the near future be at the point where frontier models are not necessary for, let's say, ninety n ninety percent of coding tasks, right? Will be able to be done by non-frontier, you know, low-cost models. And it it's an incredible development in the industry. And it's one that, you know, many of us e expected given the way that we saw, you know, the previous developments it just in computing and the internet and and and cloud computing and the rise of of of Linux and open source winning, you know, hands down in that you know, operating system and cloud market. So it makes sense that you know, would go the same direction with open source. And thank goodness because it's, you know, it's driving the cost, it's driving the token cost. The frontier model companies are at least going to have to have offerings that are that are competitive in that market. and I I and Caleb, I think what you're you know, what you're alluding to is, you know, and so we've We have, you know, for for many years, as you can probably tell, have embraced that open source. Our business has been, you know, a tremendous benefactor in the whole industry, not just Linux, but of open source you know, frameworks, you know, React and Next and things of that nature that we've used. So, you know, we're now you know, giving back to the open source community with an agentic coding framework that we've released, which is it's called Dino Stack. And you know, as I was mentioning before, where we've kind of, you know, we've encoded the the software development lifecycle into our agentic engineering practice. We're now you know making that available open source. So it was a framework that you know we developed to use internally because at the time there was nothing there was nothing like it that did you know not just multi agent, but also implemented a workflow, and it's a workflow that is dynamic based on the complexity and the risk of the work. And it includes it does planning, it will do architecture if need be. It of course does coding and will break down tasks into multiple components that can be coded by various agents. then those agents will receive a code review by a by a skeptic, and then that piece of work will also be QA'd. And this is all done in an agenc workflow. to be honest, I thought that By we developed this, it's been out in in open source for a number of months. I thought that by now it would be completely obsolete that other, you know, tools that the the frontier model providers would have released this. And while there are you know similar systems, such as ECC or agent skills, and we have some comparison, you can go to dyno stack. Umai and view the comparison to these other frameworks. there really is nothing quite like it that allows you to just clone this repo. And begin to work. And you can take advantage of that multi-step, multi-agent workflow, and then you know learn the tool and get more sophisticated with it. One of the other really important concepts that it has is this ongoing learning and improvement, right? So at the end of every task, every work session, you tell it, wrap it up and you know, encode your learning, show me what you learned, and then we're gonna write that back into the repo for not just you as the agent. But for all the rest of the team the and the team's other agents to take advantage of. So we'll s we're seeing now, you know, a r a recursive continual improvement loop happening with with agentic engineering as well, which you know, is even more of a benefit to us and of course our our customers.
Caleb Bryant: Why did you think that it was im it was valuable to make that open source to the community?
Howie W Ross: why was it valuable? like I I mean, as I mentioned, we just didn't see anything, you know, quite like it. And we we also didn't necessarily, you know, well, there was a there was
Caleb Bryant: Mm.
Howie W Ross: a potential downside that we felt, you know, we we did have an advantage by by having this system. but also, you know, as I mentioned, we had we had benefited so much and you know the productivity benefits that we had gained, we wanted to share. with the you know, with the larger community and you know, with and and also we we thought that it kinda had a relatively short you know life cycle. So we're like, we gotta put it out now. Let's let people take advantage of it. You know, use it for a short amount of time, use it as inspiration for for what is to come next. But here we are months later and we and and and many others are are continuing to use it and and get value from it.
Caleb Bryant: Great. Amazing. Well, we are almost at an hour. So we should probably wrap it up. But Howie, this is so helpful. And I'd love to have you back on as you continue to see adoption of Dino Stack. I'd love to learn more about just the evolution of of how you know that stack is improving, not just based on adoption of merchants, but also even the the community, I think. We're all in this together, realistically. And you know, the things that we can do to help one another ultimately I think will help you will help our ourselves too. but yeah, it's it's such an exciting time and I'm thankful for folks like yourself, Howie, folks like yourself, Kayla, that are that are doing this thing with me, that we can find a way to to to clear a pathway on behalf of merchants together. It really is an exciting time and I yeah, I can't thank you enough.
Howie W Ross: Well, thank you as well. Thank you so much for for the
Caila Schwartz: Yeah, thank you, Howie.
Howie W Ross: opportunity. It's a it's a real privilege to, you know, get to speak here with you with people that are so well informed on you know what's happening in the market and and and with our customers. And also, you know, just have a passion for you know, for these these customer experiences and AI and technology in general. So it's been a real pleasure. Thank you so much.
Caleb Bryant: Awesome. Thanks, Howard.
Caila Schwartz: Yeah, thank you, Howie. And thank you, Caleb. great talk as always. And I I echo the sentiments that Caleb shared and just say, you know, just to add, one thing we learned from Dreamforce was just how hungry people are for information, for context, for perspective. And so you know As Caleb said, it's an exciting time. It's an exciting time because none of us really know where this is all going to net out. and so to have voices like yours, Howie, to share, you know, just what you're seeing and what you're working on, I think is so incredibly important to the industry. So it was such a pleasure to to get to chat with you. And thank you for thank you both for putting up with my my background noise.
Howie W Ross: No no no problem at all then.
Caleb Bryant: Hey, it it wouldn't be it wouldn't be a coffee and commerce if we didn't have a screaming kid or a dino Dana in the background.
Caila Schwartz: Exactly.
Caleb Bryant: Right?
Howie W Ross: Yeah, it'll be mine next time and I and I hope there is a next time. This has been you know really great. So looking forward to speaking with you all another time.
Caleb Bryant: Amazing. Thank you.
