# Agentic Commerce Field Notes with Chuck Gahun at Forrester - Dispatch: 1 - Date: 21 September 2026 - Guest: Chuck Gahun, VP & Principal Analyst, Forrester --- **Katja Forbes** - 0:12 Hello, everybody, and welcome to Field Notes from the Agentic Economy, an interview series where we explore the new non-human layer that's landing in our businesses and lives. Today, I'm here with Chuck Gahun, Vice President and Principal Analyst at Forrester. Chuck and I have been trading notes on this for half a year, and the conversation has been coming for a while. Welcome, Chuck. **Chuck Gahun** - 0:34 Thank you for having me. **Katja Forbes** - 0:35 What set you on the path of exploring agentic commerce? How did it get on your radar? **Chuck Gahun** - 0:40 Good question. So we naturally track a lot of data at Forrester, consumer behaviour, consumer trends. We've been looking at agentic commerce's evolution before OpenAI released the ACP protocol. So we were trying to figure out if consumers are using these answer engines for product discovery or shopping or any of that, and we didn't have anything concrete. We had a hunch that it was going to go here, and then when OpenAI released the ACP protocol in, I think October of last year. Wow, it's almost been a year now. October of last year, that's when we really started tracking agentic commerce and the ability to do instant checkout on Answer Engine. That's what made it real for us. **Katja Forbes** - 1:22 Was some of it a little bit coming from the clients that you were talking to, or did you feel like it was more instigated on the Forrester side? **Chuck Gahun** - 1:30 Neither. I think it was more instigated on big tech side. That's probably the best way to put it. OpenAI and then Google's announcement in January really set it all off, and that's what created the flurry of Forrester clients reaching out to us and saying, "Hey, what should we be doing about agentic commerce? How should we be thinking about this? What does this mean for our business, our commerce operators, our website, all of that? **Katja Forbes** - 1:50 You've recently published the report "AI agents are your new target audience", drawing on contributors from all around the world. Forrester landed on a fairly deliberately unglamorous verdict there, which is agentic commerce is advancing but not yet autonomous. So, in a market where every vendor is selling inevitability, why was it important to sort of slow that story down? **Chuck Gahun** - 2:16 So, let's start with some data points. Right, 38% of consumers are using answer engines to discover new products, and but only 17% of them are actually completing a transaction in an answer engine. FYI, significant **Katja Forbes** - 2:27 gap. This **Chuck Gahun** - 2:28 is like March, April data of this year, so not that outdated. And if you look further beyond that, you see that if you take commerce out of the equation, only 13% of U.S. and U.K. adults are even okay with allowing an agent to complete a task on their behalf. So the metrics are pretty low here. So when we started to look at this, we said, okay, let's say that these are US UK numbers. Let's hold it to this part of the world. Let's go global. Let's look at let's look at China. Let's look at Japan. Let's look at other parts of the world and say, hey, is there autonomous agentic transactions? Are they happening out there? And we looked far and wide, and honestly, we didn't find a tremendous amount of data or stories that this was happening. We found some examples with Alibaba and Alipay, that whole side with like some semi-autonomous transactions that had started to happen, you could order food as part of the whole, you know, almost like imagine if you were on Amazon and you could order like food takeout as well. You task an agent to go do it, meet meeting the criteria, meeting your meal preferences, dietary restrictions, all that. You can probably get the picture there. When we were autonomous, we found a few actually intelligent factories that were using B2 storyline for agentic autonomous transactions, but on the consumer side, they were just really hard to find. So that's really how the narrative landed in that we are much more trending towards semi autonomous right now, not autonomous. **Katja Forbes** - 4:00 Yeah, I mean, some of the stats that I've read as well is that the B2B side is where the immediate opportunities are. So I'm not surprised to hear that that was where you were seeing a bit more of that trajectory. **Chuck Gahun** - 4:12 And there's, I mean, there's a lot more advanced thinking and guardrails and protocols for B2B payments as well. So really, lends that side of the house advancing. I would say for the first time I've seen in my career advancing faster than B2C. **Katja Forbes** - 4:27 I'm just going to pull a couple of threads there. What do you think is the the reason why B2B is able to advance faster than B2C in this moment for agentic commerce, **Chuck Gahun** - 4:42 well, answer engines are very good at providing contextual responses and at least contextualising the content when you have a question, right? And so, when you think about a B2B buying group, you can think about how many different players are involved there, and they're all looking for different bits of information. The CFO versus a procurement versus a business person versus a marketer, and so these answer engines have started to at least in the earlier stages of where they're you're doing research and you're evaluating options. These answer engines are helping these buying groups in B 2b orgs advance faster, make decisions at least maybe even get down to like if I'm looking for spare parts for a specific type of you know fleet of vans. I have I can get down to the manufacturers of those parts faster, reach out to them rather than doing this endless web research. And so, in some ways, and one of the in state of agentic commerce which we published in I think early June, it released. That's one of the notions that we landed on as well, which is that as B2B advances here, which is I think very very interesting. What's actually going to happen is that over time the pace of business will increase. When a business to business interaction starts to get faster, more intelligent, all of that, we will all do business in a little bit more of an optimised fashion, which is very interesting to analyse. **Katja Forbes** - 6:06 Yeah, it's hugely interesting, and I think a lot of businesses are very interested in that optimization narrative for the way that they are doing things. But you have said that the hype is running ahead of the behaviour, and you know, as we've read and seen, narratives assume agent buying is everywhere, which from your data is not actually accurate. But the trust and the adoption is definitely still uneven. So where's the widest gap between what the market believes and what your data actually shows? **Chuck Gahun** - 6:36 Oh, it's that it's 100% payments. So yeah, we so we are really in agents shopping world. Just today, I was talking. I was interviewing somebody else for another piece of research, and Gen Xers said the same thing. Even I've even speak to spoken to Gen Z folks that I've been interviewed on these things, and the storyline is the same thing, which is I will use the answer engine to find options, to do discovery to understand what products make sense, why contextualise responses back and forth. Then, when it comes to making the transaction happen, I'm going to open a different tab in a different browser, and I'm going to go and make go to the site, the brand site, or the retailer site, and execute. So the payments is, and the reason why our data shows is like consumers dislike the loss of control. They worry about something going wrong. In fact, in the state of Agentec, we published that our data shows that 44% of consumers would hold the answer engine responsible if something goes wrong with the commerce task. **Katja Forbes** - 7:34 The answer engine-that's an interesting finding. **Chuck Gahun** - 7:37 Or Google-they would hold them responsible, not the brand together if they transact through one of those two channels. So it's interesting thinking about from a consumer lens, right? **Katja Forbes** - 7:48 The report, the State of Agentic Commerce report, it frames the brand advantage around three levers in every agent transaction. So the levers that in the report are context, fluency, and trust. So, can you walk me through why those three and not things like schema and visibility checklists that everybody else is selling? Oh, yeah, good question. So, the reason why the AI agents or your new target audience even came about was I started asking questions, and I believe like February or March of this year, and I started asking leaders and services and software vendors. I said, "Okay, let's assume that I have made my site visible to AI agents. **Chuck Gahun** - 8:30 Let's assume that I've done that. What now? And the response was, "Well, what do you mean, what now? I'm like, "Well, how do I actually make money off this channel? How do I convert? How do I engage consumers more? How do I maybe just have a better customer experience with my customers, human customers? Any of that in play here? And people didn't really have the answers until I started getting deeper down the content strategy rabbit hole. And as I went further and further down, I realised that what's happening here is as a massive exercise in operations and commerce operations and content strategies really unfolding here. And as you start thinking about that, you say, "Okay, if an AI agent was a persona, how and what would this persona be looking for? What would what would be the interests here, right? And you start to realise, well, first of all, if I'm an agent and I'm coming to your side, I need to make sure that I have enough context so that when a human is asking me questions, I am resourceful to the human, right? So in that motion, I need to have context and content from owned and unowned sources. So you get the context framing. The fluency piece was interesting and how it came about in many many conversations, and this one is really focused on you know just as you have design systems and content taxonomies, we can orient humans to better interpret websites. Agents are also looking for fluency in the brand, in the brand narrative, product terminology. Right, they're looking for. Standard nomenclature. It's not they're they're coming in, they're retrieving, and they're leaving. This isn't about browsing. This isn't, and that's what has marketers rattled because this non-deterministic experience. Your navigation flows that it matters very little to an agent. Like you're not going to take an agent down a specific path the way that you want to guide them, like you would through a digital marketing campaign, right? And so, when they take this content away, they need it to be fluent. And then, if you think about the and the fluency also has to do with, like, if I'm an agent, if I'm a good agent, I don't want to infer and hallucinate because the human that's going to read the stuff that I create is going to lose trust in me. So, how do I? The third one, which is trust, is all about how do I deliver consistent, verifiable facts. How do I make sure my information is up to date? And that's where the third leg of the stool came about. So it was context, fluency, and trust. **Katja Forbes** - 10:55 Yeah, and trust is just so crucial to any of this actually functioning, both in a B2B and a B2C space, so I think fluency was the one. I remember when we talk when we were talking about this, fluency was the one that kind of needed the most unpacking content that AI system can actually understand, that it can use with confidence rather than just find. You know, when we were kicking the ideas around earlier this year, the concept took a while to click, even in like my brain. How did it earn that place as the distinct lever? **Chuck Gahun** - 11:28 We we look far and wide. I said, you know, fluency seems a little counterintuitive, but and I asked a bunch of my peer analysts. I said, help me find a better word, and we kept coming back to that. It has to be fluent in what it's taking, your site has to be fluent, and the agent has to be fluent in the the content. And it was really for two reasons. One I mentioned before, which is around nomenclature, right? Like if you are talking about everybody is talking about running shoes, and you are calling them running sneakers, it's gonna it's not gonna know when to surface your product in a response to a human query, right? And that's that's a problem. And then if you look at the other side, it's you know we've been talking about hallucinations and content and answer engines for this contextualised information for a while, but they want to reduce inference, meaning they don't want to infer. So if they have to create the reality of your site that is not good for your brand or good for the agent, and that's and there's of course there's more processing power buying and all that. Sure, but more importantly, when you're thinking about the content strategy here, and so when we looked at both of those levers, you know, reducing inference and nomenclature being per market standard, we realised fluency was the word that fit that the best. **Katja Forbes** - 12:46 Yeah. Well, beyond fluency, when you brought those outside voices into this work, because there's a number of people who've contributed to the report, what perspectives change the shape of the report? What did you get from those conversations that perhaps you wouldn't have reached inside Forrester alone. **Chuck Gahun** - 13:03 Yeah, yeah, yeah. So I'll give you two or three that were really important. One was this notion of a horizontal versus vertical content strategy, and and where is the white space? Meaning, when you think about, are you trying to own a product vertical, like do you want to have content strategy in running shoes, or do you want to have a content strategy in shoes and how to select the right type of shoe for what occasion? Like this is these are two fundamentally different content operation tasks. But for an agent, you can think about how this means whether your brand and your products are actually something that they they can reference even in other parts and other prompts that might surface, or just specific questions about running shoes, right? It's neither right nor wrong. It depends upon every company's strategy and what they want to do. But the but the nuance of that came through when I was talking to Harmon International, you know JBL speakers, and we were talking about the white space of content strategy, which is everybody is going to have you know what are the best headphones for running, great. How many are going to have content on what are the best or the longest lasting headphones that are sweat wicking. Good question, right? So, what is the white space of content strategy that you're trying to own as we're opening up new channels, being agentic channels? And let's be real, what's really happening with this agentic, these agentic channels is that it's taking that piece of digital commerce and it's fragmenting it further. Right now, you have digital commerce, and then you have agentic commerce as a part of the storyline. So that's that was one. Number two was we found you know context markdown files. Everybody's talking about it. The more and more we dug into it, we found that there was really no evidence that agents are digesting this. So we put. Able to largely experimental, that was one thing, and then the last thing that was very interesting was we were thinking about forward-looking ideas, agentic channel loyalty, even like what content goes out on agentic channels versus not. That was another notion that came about. It was interesting because everybody's like, get the schema going, make sure your content on all your all your content on your website is visible. Well, what if that you don't want to do that? What if you want only want to have some content available? I mean, publishers are thinking about this heavily right now, and that's been one of the challenges in the you know the news and media publishing world. So those are the three that come to mind as you ask that question. **Katja Forbes** - 15:37 Yeah. So I think the the analysis that I've been doing since February, with all of the articles that I've been collecting, I have a research agent. It researches for me every day. I have a database of more than 1000 articles on different aspects of agentic commerce or AI agents in the ecosystem, the agentic economy. And when I've done an analysis of that, I've found probably about 70% of that narrative and conversation is around it. It's around the rails, the the plumbing, the utility, the pipes, and people are really starving for for people are really starving for strategic conversation here. I guess it's a little bit because that's what they can sell. They can sell rails and pipes. **Chuck Gahun** - 16:26 Yeah, yeah, very true. Very, very true. It's a good way to look at it. **Katja Forbes** - 16:30 Yeah. So and so since it went out, the report. What's what's been the reaction from all of the people that you most wanted to reach? You know, the the commerce leaders, the CX leaders. You know, clients who are pressure testing it against their own roadmaps. **Chuck Gahun** - 16:44 I have I have never published piece of research that has had more demand. Sincerely, **Katja Forbes** - 16:48 that's so fantastic. **Chuck Gahun** - 16:50 This AI agents are your target audience. I've done so many client guidance sessions and deliver speeches. It tends to be the topic this year that everybody is focused on. Like, how do I actually not make my content visible to agents. How do I actually target an agent to get the desired behaviour that I want? You know, you have like the evertunes and the profounds of the world that are doing this work at a high premium and a cost that many businesses might not be able to achieve afford, right? And so when you start about it, you say, okay, like what is the content strategy to make this palatable for most businesses? And workshops, actually sitting down, working with clients, looking at their content, helping them think through like FAQs and whatnot. So to make it more digestible, smaller chunks of content. That's it's been very interesting, very very interesting. **Katja Forbes** - 17:37 Yeah, that's an amazing reaction. You must be super pleased with that. **Chuck Gahun** - 17:40 Yeah, yeah, absolutely. But now you know, on and off to the next **Katja Forbes** - 17:44 one. Yeah, I know, I know. So let's let's talk about a little bit of that forward-looking stuff. So the barrier that you've named in the report, it's not technology. It is some of that internal silos, the immovable cultures, and it reframes the agentic commerce from being an IT project to being something that's more of an organisational design problem. When I'm talking to leaders as well, I'm seeing a real need for mindset shift in this. So, which are the teams that are furthest away from realising that that's the fight they're actually in, or that's the challenge that they're trying to answer. **Chuck Gahun** - 18:22 Yeah, this is not a tech issue. This is definitely a content and data issue. And the teams that are the furthest away, honestly, that I've spoken to are the ones that, well, a few of them don't really believe that this is going to be reality. Which so if you move beyond the sort of the naysayers of it, then you start moving into a space where you realise that some organisations, their content processes and content ops, over the course of years have created so much siloed content that the idea that they would ever be able to wrangle their head around all of the published content out there, think like outdated marketing campaigns. Think about inaccurate product descriptions on on like like pages that are just you know standalone pages from old marketing campaigns and landing pages and stuff. Those are the businesses I think that are the furthest away because these agents and training crawlers are crawling the entire World Wide Web, and that they're not just crawling, you know, your most recent published site. So that's yeah, **Katja Forbes** - 19:30 and and it's I guess some of them are kind of shying away from the fact that there's a really huge amount of boring cleanup work that has to be done in order for them to get to this exciting agentic future. **Chuck Gahun** - 19:44 Big, big, big, big opportunity for commerce. This is all commerce operations exercise, is it not? Commerce **Katja Forbes** - 19:52 operations is that a new term? **Chuck Gahun** - 19:54 No, no. It's just like when you think about what's really happening here. The you know the upcoming. Agentic web is preparing people for getting their docs in a row ahead of time for the AI agents' target audience. And we spoke to a product leader who said that they released p1 in market and it was an abysmal failure. Like it didn't work. It was a spare part story. The whole product was a flop. Well, of course, Reddit picked it up and all the reviews and they were being completely bashed, and so what they did was, as they released like p3, p4, the product got better and better. What they did was, they took their content and targeted it directly at Reddit, pulled it into their narrative, told the p1 story, and then told the rest of the story. And guess what? Two months later, you go to an answer engine and you type in information about that product, and it tells you the full story now. **Katja Forbes** - 20:45 Oh **Chuck Gahun** - 20:46 That is targeted content strategy. Yeah, **Katja Forbes** - 20:49 yeah. **Chuck Gahun** - 20:49 Off their it's all their pain point where consumers now understand that, or buyers, B 2b buyers understand that the p1 of the product was not as good, but three and four are so much better, and the here are the reviews for it, and they get the full story, not just the bad Reddit reviews from you know the first one. But I heard as of this weekend that now it's not indexing Reddit as much. So I don't know. We will see where this all goes. **Katja Forbes** - 21:12 Yeah, this is a this is a part of it. Like you know, Agentic Commerce is weird and unpredictable, and what happened last week with it will probably be different with what happens this week with it, or what happened yesterday might be different to what happens today. So it's really unpredictable, and I can understand, you know, the nervousness in businesses trying to wrangle this kind of stuff. **Chuck Gahun** - 21:34 Yeah, yeah, exactly. Very tough. It's it's constantly shifting, and they don't. By the time they figured out what is what, there's a whole new model release. There's a whole new like the the AI race is just challenging businesses and in their agility. **Katja Forbes** - 21:50 Let's lift it up out of the tactical into the more strategic because there is a control question underneath all of this. Yeah, which is you know who sets the terms when an external agent shows up to transact? Because businesses have spent like decades curating how they discovered, and that control is definitely slipping. When I talk about this, I call it commercial sovereignty. Those terms under which a business chooses to engage with an AI agent, somebody else's AI agent, or even a robot, a car, a drone, or other autonomous actors that can actually transact. What do you think companies should think about in terms of keeping sovereignty over its own customer relationships in this context? **Chuck Gahun** - 22:29 First of all, clients have not asked us about this yet, but I get the sense it's coming, and the reason it's coming is because I have started talking to a few clients about how they should architect, hence some of the research I'm doing. How they should architect their agentic layer, and how should they be thinking about the customer experience? I think the next stop on the train is the policies, the guardrails, and a little bit of what you're talking about, which is what are the terms with which they are engaging, interacting, transacting in the agentic world, and when when are they playing, and when are they not playing, and what does that look like in an agent agent interaction? I think that that's right on. I would say definitely by this time next year, this will be a hot topic of conversation. **Katja Forbes** - 23:16 Yeah, I find when I talk to people about it, they've not really thought this far ahead, and they're still, as I say, because a lot of the narrative sits at that foundational tactical level, and so being fed that this is what you need to think about means that the conversations are not happening at that super strategic level. So, I mean, you just talked about that you have the architect's guide to agentic commerce coming? What questions are you answering in that guide that you didn't answer or deliberately left open out of the last report? **Chuck Gahun** - 23:51 So the first piece we did was state of agentic commerce. That was more about data, consumer trends, even B2B trends, like what is happening with Agentic Commerce. The next one we did was AI agents to target audience. We also did a few other, like B 2b payments, state of B2B payments, those types of things. And when we looked at AI agent target audience, that was content strategy. This architect's guide is looking at the tech and going to provide a more robust set of guiding principles, guidelines. I don't want to use the word guardrails because that obviously means something different here. But things like payment protocols, existing commerce tech, what should you do? What happens to things like order management, inventory management when you know the ACP in OpenAI expects doesn't expect, but provides you the option to refresh every 15 minutes? Like, what does that mean to API performance? It's going to set out the tone for what leaders should be thinking about. Our guidance has always been like: first, figure out if you want to sell on agentic channels. If you do want to sell on agentic channels, then there will be this architect's guide to guide them through on how to actually build that tech infrastructure, even commerce process. Like, how often do you update your PDP? I was just talking to a client last week who told me that they update their PDPs twice a year. Is that going to apply in the agentic commerce world? Depends upon your product, I guess. But I don't know. Probably not. **Katja Forbes** - 25:13 Last question for you: If a CEO listens to this and does exactly one thing differently on Monday, what is it? **Chuck Gahun** - 25:24 I'm not going to give the age-old advice of not to wait, but I will say be very deliberate about what I'm calling how you're going to colonise the agentic space. Think about how your brand shows up, when it shows up, where it shows up, and why is it showing up and why not? Think that build a strategy for your business on these agentic channels, and it might not be your entire product suite, it might not be your entire set of services, but do track the fact that consumer adoption is starting to accelerate and it's lagging. Consumers don't trust agentic payments, but discovery, shopping, experiencing brands, understanding brands. Heck, I'm even starting to talk to customers a lot about moving some of the functionality over into agentic channels like WebMCP. So, as some of the pieces start to move out of websites into agentic channels, what is you know, Mister or Misses CEO, what is your strategy to drive your business through that channel? Is be deliberate about it. **Katja Forbes** - 26:27 Yeah, the intentionality is really crucial because this market is real and this market is scaling fast. **Chuck Gahun** - 26:34 One of the other things I will share is that we are working on two big pieces of research. While I'm that I'm leading one is this agentic architect's guide for agentic commerce, and then the other one that we're also looking into is agentic experiences. And I'm leading that with another one of my colleagues as well across the pond and in London. And we are talking about things like A2 UI. Will these agents actually exchange design elements, user interfaces? Also, like, how do you engineer an agentic experience? So, lots more fun stuff coming up on this. **Katja Forbes** - 27:06 Yeah, I'm interested to see how the generative UI picture plays out in in that space where you know the the fixed customer journey ceases to be a thing, and we now have you know interfaces generating on the fly, depending on whatever your query was, whatever you're trying to do, and that's both for a B2C and a B2B. **Chuck Gahun** - 27:27 Yeah, yeah, yeah, absolutely. That's and I think that's going to be very, very interesting as it as it unfolds, because because the generative site is going to fundamentally change our digital experiences as we know them now, and I think, in some ways, will create the pathway for agentic channels in in its own, which which is kind of like the whole AI labelling watermarking thing, right? Once you see that all this content is AI labelled or watermarked as a human, we're going to say, well, where's the human content? It's going to draw your eye to the other side of the equation because you know what is already. **Katja Forbes** - 28:03 Well, thank you for all of that. That was fascinating. It was really great to connect on the work that's been done, the work that's coming up, and I appreciate you giving me time on this. **Chuck Gahun** - 28:11 You got it. You got it. Always good to connect. Thanks for having me. **Katja Forbes** - 28:15 Thanks, Chuck.