Episode 210: How AI is rewriting the rules of content and brand strategy - B2B Marketing Podcast

Meet Dan Miles: Executive VP at The Hoffman Agency Hello everyone and welcome to the B2B Marketing Podcast in association with the CC Group, A Hoffman Agency. My name is Kavita Singh. I'm Head of Growth Solutions Content and today we are going to be discussing how AI is affecting the way B2B marketers are approaching their content and brand strategies. We are joined by Dan Miles, the Executive Vice President over at the Hoffman Agency. Welcome, Dan. How are you doing today? Speaker 2 Thank you. I'm I'm very well. How are you? Speaker 1 Yeah, doing great. I guess to before we kind of get started on the topic, which I'm really interested in diving into today, can you introduce yourself and tell me a little bit about your role? We're in depth. Speaker 2 Yes, of course. Thank you. So Executive Vice President at Hoffman, Hoffman Europe, CC Group, just as a a point of of clarification, CC Group was an agency that was acquired by Hoffman last March and we're in that process of still a legal entity as an agency. CC Group as well as, as well as, you know, being part of Hoffman, my role as Executive Vice President of Marketing Services, specifically looks at integrated marketing programs. So I work with technology brands that want to take a more multi channel kind of multi format and through the funnel approach to to achieving their business objectives. Hoffmann are an NCC group are traditionally A comms agency. So you know, kind of public affairs strategy, PR and and analyst relations. But increasingly PR and comms is a, you know, it's a multi channel consideration. And and so it should be, you know, the way that tech buyers buy is evolving and is a is a multi channel consideration. So, yeah, those those organizations that that want to think about how they can tell their story, how they can disseminate their thought leadership, how they can be brand LED everywhere where their buyers are and consume information comes through me and my team so well. Speaker 1 It's great to have a little bit of background and perspective considering today's topic for today. So great. Yeah. Thanks for coming on the podcast. I guess to kind of get started, if we compare how buyers research solutions three years ago versus today, in your opinion, what do you think is sort of the biggest difference there? From Linear to Delegated: The New Buyer Research Journey Yeah. I think that's the that's the crux of it all really. Three years ago, the buyer journey typically and I think we can talk across B to B markets, but specifically B to B.Tech was linear and manual. I think that's, yeah, I've been thinking about words to use describe the shift and it feels like we've gone from something that was linear and manual in that a user would type a keyword into, you know, a search engine or, or would use a sequence of keywords wherever they happen to be in order to be able to open 5 or 10 tabs. And then manually would synthesise the search results and kind of synthesise the information across different tabs that were open in order to form themselves a comparison in terms of what might be the most appropriate source of information for them. So that kind of linear and manual process was the only kind of automated component of that was the the retrieval of that information. I think today the journey is a little bit more latent or, or kind of tacit and delegated in that a user or or a buyer asks a prompt. There's a, you know, I think we're all aware of, of a linguistic difference between a traditional search engine kind of keyword search and a prompt. But whatever happened, whatever we happen to use as a prompt and LLN will then perform that synthesis of information for them. There is no need and there is no opening of five to 10 tabs. There is no manual process of reading different bits of information from different websites and coming to a conclusion yourself. The LLM and kind of generative AI certainly performs synthesis for them. The kind of tab open phase is disappearing. Research has shifted from finding the right link to receiving a verified answer. Yes, I think we're still in the phase where we are doing some background checks on the citations that AI uses in order to provide that synthesis. But it's very rare that we, if we do look into the back end of how and why an LLM has provided a recommendation for us that we disagree with the citations they use. And we'll go into over the course of a discussion how LLM's use information in order to make those recommendations. But it's it's less linear and less manual and more delegated and you know, the the emphasis of on the human has shifted from finding the right link to reading the answer, kind of receiving the the summary and the verified answer. Speaker 1 Yeah. And I'm glad you brought up, you know, even though it does feel like you're getting an instant answer, there might be some sense checking there, but you're kind of already down path that's sort of been determined already, which I think is interesting. Do you think we're officially in sort of this zero click era and you know, when we talk about organic traffic, you know, is that basically over because I think there might be some misconceptions there. Speaker 2 Yeah, I do. I think the use of zero click has become a a catch all for, for the end of things that aren't really ending. I do think that we are in we're officially and you know, objectively in a zero click era as far as information queries is concerned. If, if there's AB2B brand, if your content exists simply to answer what is X or how does my product do this thing? Your, your organic traffic is going to vanish because then then can provide that answer, you know, without a click as a response to a prompt, that kind of binary, almost simplistic question and expecting a simplistic answer, that type of of kind of informational query. There is no, there is no need for a click on that. And I think the zero click here is, is certainly over in that respect. However, organic traffic is not over and SEO and the, and the fundamentals of SEO are certainly not over. I think organic traffic is being redefined. We, I think we're all seeing reductions in organic traffic because a lot of that basic informational request is being dealt with by, by LLMS and they will only get better at doing that, let alone agents. And we'll talk later about the difference between the generative and LLMS and, and kind of a genetic AI, but traffic will still exist, but it'll be rarer. But higher intent If we, if we think about the fact that informational queries and almost basic comparisons that the early stage of a of AB2B kind of search journey or buyer journey, top line comparisons and general generic prompts information gathering, that's all going to be a zero click. So when an individual does get to our website, naturally they've already been through that process. So there is innately higher intent in that the goal is no longer to get the click for the answer. The goal for brands is to be the cited source that the user clicks on when they want to go deeper, when they actually want to for whatever reason, get more information than AI summary can provide. That is innately higher intense. And the reality is from an SEO perspective that the the fundamentals of SEO retrieval in terms of signals, website signals, domain authority, the basics of how an LLM brings together lots and lots and lots of different sources in order to synthesize that information. Those fundamentals of SEO are still exist. So it is still important to think about signals and SEO retrieval. It's the new paradigm that LLMS provide is that you need to be thinking much more broadly about where your information exists and why it exists, what kind of queries it answers in order to allow LLMS to use multiple sources to synthesize the same kind of facts. And that's the kind of the difference for me that the zero click era that we that we're in that describes this kind of era. It's about thinking in terms of multiple sources. Discovery rather than awareness is something we're talking about internally, Hoffman. Discovery means understanding the third parties, media, social high, high domain authority and high authority individuals, as well as your own website are critical, a critical part of content strategy or a critical part of providing LNMS with what they need. And the clicks you will be lower, but actually they'll be more valuable. And that's understanding that I think it's a really important way to to understand the role that websites will have going forward. Speaker 1 Yeah, definitely. I think in order to, to sort of come up with a strategy that meets where we're at today, it's actually really important to understand exactly what you just said and, and why that is. Understanding How AI Selects and Synthesizes Information I guess with that being said, you know, if buyers are no longer clicking in the same way, what does AI actually look for when it decides, you know, what to surface? Because I can imagine a lot of people listening would probably want to know the key to that. Speaker 2 Yeah, I think it's a really interestingly worded question. It's quite natural for us all the way that we use or a lot of us use LLMS and generative AI. It's like an always on to and fro of a conversation. So the the kind of anthropomorphizing of AI is understandable. So words like what or phrases like what does it look for is understandable but actually not technically accurate in terms of information retrieval or synthesis. And AI doesn't rank sites, it doesn't look for anything. It is a probability engine. It simply synthesizes lots of information in order to provide the most statistically likely completion of a prompt. Another kind of concept and phrase that that we are with that we're thinking about is marketing in a probabilistic age and and that SEO isn't probabilistic. SEO is, is largely a volume play. But if by calculating consensus and looking for as many different authoritative citations to synthesize, it kind of looks for three core things, entity association. So does the web, Reddit, LinkedIn, journals, media, particularly media and as well as individual websites and and social. Do those parts of the web associate your brand with a specific attribute or expertise? And that consistency is really, really important. And as I was alluding to in terms of value expressing a value proposition is a multi channel discipline now. So entity association is really, really important across multiple different types of of parts of the of the Internet. And secondly, cross web verification sometimes known as the rule of four. But what this means really is that if the AICS you mentioned on your own site, that is a claim, not a fact. And the different understanding the difference between a claim and a fact is really important in, in driving and improving your AI visibility and, and really understanding the, the difference between awareness and visibility. What you say about yourself, you know, probabilistic aid in a zero click era is at best academic at best, because an LLM is not influenced by a claim. But if it sees the this same, the thing, the same thing you say about yourself is also corroborated on a site on a tier one journal and on a Reddit thread on Wikipedia, increasingly on the open parts of of LinkedIn. And that and LinkedIn is one of the most increasingly influential sources for for LM citations. That moves from being a claim that you've made on your own site to a fact that is corroborated across multiple sites. And so you've moved from simply making a claim about yourself to there being corroborated multi kind of corroborated facts about you. So this cross web verification is critical. And then thirdly, thinking about the way that you talk about yourself and providing content in as kind of citation based, So thinking about extractable statistics, you know, unique frameworks using attributes that are widely used and understood is really important. So I think the days of of category creation are over. If you want to create a new way of describing your proposition, you are not giving LLMS what they need because what they need is the same attribute, the same word, the same concepts repeated across multiple different types of website in order for it to see it as a fact. If you want to invent a new word to describe your category, good luck. Like you've, you've just chosen to not be visible to LLMS. So extractable insights, unique frameworks, similar concepts, but unique ways of of framing that concept and using expert quotes. These are easy for LLMS to kind of snip and repeat and offer to users as a completion of a prompt. These are the things that drive the drive visibility and improve AI ranking or whatever term you want to use in order to to think about this new age of discovery. Why Strong Brands Can Be Invisible to Generative AI Yeah, that's really interesting. You know, how can a brand be strong in the market yet completely invisible to AI? Or is that even something that could happen? Because I'm curious. Someone's listening and they think they're positioning as a brand is strong, but they haven't even thought about the AI aspect of it. What does that look like? Speaker 2 Yeah, I think it's about silos again, the brand graph or you know, the AI graph. This is a concept that that people talk about a lot in terms of the amount of information that needs to be consistent about your brand across lots of different types of platform. If you don't have that, if you're not part of the graph, you don't exist to an L&M. So I think that is possible for brands to be to to be conventionally massive, but from an AI in an AI context, invisible. If all of your best thinking and if all of your thought leadership, if, if the way that you bring your value proposition to life is locked inside gated PDFs, please stop that by the way, or or behind login screens or, or if you do all of your best thinking in your most expansive description of your value proposition on your own website and you don't and you aren't thinking in multiple channels about saying the same thing repeatedly. The AI is no way to kind of verify your authority. All you are is making a bunch of claims. You are not turning a claim into a fact. I was speaking to an SEO friend last night about, you know, about this interplay. And I think there's another way of of being less visible to to AIS and that's not thinking about or only thinking about conventional SEO in terms of keywords and long form content. If if all you do is is write your content for human consumption and you don't think about LLMS needs facts. So they need crossword verification, they need distribution agents need more than facts. They need to be actually to to be able to use your information to complete a task which needs a specific type of structure. So if you're not structuring for agentic and and distributing for LLMS, but you're still writing lots of content, you all you're doing is writing for human consumption. And I think that will increasingly be a factor behind low LLM visibility. So those are kind of really important considerations and they're those are ways that brands can feel like they're big. They feel like they're doing the right thing in terms of lots of content, thinking about the leadership, thinking about brand and product LED stuff. All of the kind of stuff that we've been as marked as we've been telling each other for a decade that actually there's a new paradigm to this now. Speaker 1 Yeah, no, definitely. I think that's really interesting. You know, off the back of that, you know, if someone's approaching this exact problem, is it, you know, is it something that could be attributed to like, oh, it's an SEO problem or there's an issue with the messaging and positioning? Or is it kind of just a combination of all of those and kind of getting a balance there? Speaker 2 Yeah, I think it's almost sequential. I don't think that. I don't think it's, it's an either or from in terms of, you know, LLN visibility versus SEOI think it's important to be decoupled to decouple generative visibility versus agentic usability. And, and this year, particularly with things like open clock, this rapid adoption and use of of open source agents like like open Claw. And I think this year is going to be a big year for the increasing use of agentic AI. And it's really important to distinguish between those two LLN visibility, which we've had for a couple of years and which is still, you know, increasingly influential is a messaging and positioning play. You need to win the consensus by ensuring your brand's unique point of view and your attributes are echoed across those influential parts of the web. So that is a kind of a messaging problem. But in in order to make your particular, your website, but also all of those influential parts of the web consistent, you have to adhere to basic SEO kind of principles. And SEO has been saying for a long time, the more you make you make content native to different platforms, the more authoritative your your domain is going to be that will remain the same. So, so step one is always think about content from an SEO perspective, but the degenerative visibility component to this is about echoing those consistent attributes, those consistent point of views, those unique frameworks across lots of different parts of the web. A genetic usability is about structure, code and ensuring that an agent can programmatically navigate your site to complete a task. And I've seen brands go already this year, go a long way to making their content fit for a genetic usability by actually stripping pages back just to a site map just to the code and actually which renders it largely illegible to a human entirely. And so foregoing the human comprehension side and, and only providing information for agents. And I think that's probably too far for now, but it does raise the question of what is a website going to be in, in a very, very near future, because I think that is very much subject to change. But I don't think it's an either or. It's not an SEO problem. SEO is part of the of the solution. It's it's step one in terms of basics, but that isn't job done and decoupling generative visibility from agentic usability. I think brands that are going to be ahead gonna be doing that this year. The Critical Difference Between AI Summaries and Agent Actions No, definitely makes sense. I mean, you touched a little bit on it already, but to dive a little bit deeper, what changes when we move from AI summarizing content to AI actually acting on behalf of the user? Speaker 2 Yeah. It's so interesting and I think it's important at this point actually to Kathy, everything that we've discussed in context of we are where everybody is still learning and there are no definitive right or wrong answers. There are lots and lots of potential right answers. And you know, I certainly feel like I'm like I did in, in the early 2000s where I'm learning this new, this new dimensions of marketing. And so, you know, the descriptions that I'm giving are my current understanding of where we're at. And my excitement about what you know, what that means in terms of new ways to think about it is certainly not definitive. But the way I think about it is when an AI summarizes, which is typically, you know, an LMC, it's a librarian, right? All it needs to do is go and find as many different parts of the web that corroborate the same fact. It just needs to find a book. It's retrieval in its conventional sense, retrieval of information and then synthesis of that information as a user experience. When an agent acts, it is acting as an assistant and it needs to use the information to complete a task. It is not simply synthesizing information to provide a summary. It's been given a task and it needs that information to be able to complete a task Summary pricing requires authority. And you know, this kind of echo of information. Agents require determinism. And in the B2 marketing context, one of the best use cases for a genetic I can I've seen over the last few months is procurement. So procurement is one of the biggest adopters of agents, procurement agents. So often in the first stages of an RFP, in the first stages of procurement process, an agent will be tasked with going out to find a series of vendors have this technology, have this entry level price point, have these, these levels of compliance or ISO regulation or, you know, geographical legislative adherence. And that the the agent where they'll go and go off and then provide us a complete that task, right. Provide a complete rundown of which vendors meet that criteria. That's deterministic. The agent needs to know that the the book of demo button will always be in a specific place in order to be able to know that vendor is easy to buy from it. It needs the information that it gathers to be more than just a, a description of a value proposition. It it needs to be able to pass information and use that to answer the question, does it, is it compliant in this area? Does it mean are the entry level prices this and all between this parameter and all of that information needs to be defined in a in Jason, right in a specific language and clear and not hidden behind a unique language or or a different way of describing a category or a, you know, a clever kind of marketing pun that deterministic and that content structure is, I think, the difference between summarising and acting. How to Spot If Your Brand is Invisible to AI I think it's really important to understand that sort of distinction between the two. You know, I think it'd be great to know what what do you think are the first signs that a company is already invisible in generative outputs at the moment? Speaker 2 Oh, it's a great question. And I think first and foremost, it would be remiss of me to not say that there are tools out there that can provide a, a very useful perspective on your LLM visibility. We have one, it's called Jedi and there are lots of others. Semrush have a good one that I use sometimes as well. And I would say the first port of call before you need to think about reviewing your own content and you're reviewing your own multi channel comms in order to make some kind of assumption about whether you are visible. Use a tool. And again, I think the the excitement right now is that there are lots of different tools offering interestingly different ways of doing things. But don't expect, don't enter it into this process of trying to find out if you are visible and why. Don't enter process expecting as kind of definitive answers where there are lots of different ways of thinking about it. But beyond the the use of tools, there are a few red flags. And I think the big one is the kind of generic recommendation. If an AI is asked for a solution in your niche and the AI lists 3 competitors and then says there are many other providers. What that means is that they're the three competitors are doing a better job than you are of running through those three. You know, really important components of LLM visibility, entity association, crossword verification, and citation bait. If you run a series of prompts that you would want to be searched, surfaced for by an LLM and you don't feature in those, but you might be listed in other providers, then you're not giving the LLM what it needs in order to be able to be visible. There's another big concept around AI, which is this concept of hallucination. So, and, and in, in the context of signs of invisibility, brand hallucination is a case of kind of the AI. The AI would attribute your, your framework or your products to a larger or more visible competitor. That happens a lot. So there's a really interesting tension between using concepts to describe your category and describe your attribute that are recognised enough for an element to say, OK, that's a fact. That's really important. So there is an element of use of the same kind of language to describe on a top level, to describe, you know, product or service or a solution. It's how you then contextualise that in outcomes and in news cases and in frameworks, and then how you echo that across, you know, the important parts of the web and how consistent that is. That is the differentiation between good visibility and bad visibility. And if you don't have that, and we find this a lot, what you might find is LLM is attributing your value proposition or the way that you describe your value proposition or your, even your products to a larger, more visible competitor. That does happen. And that's a, that's another important sign. And then I would say footnotes, 0 footnotes. If, if you see your direct traffic staying steady, but your AI referral traffic, you know, in GPC or perplexity or cloud, whatever is non existent in terms of referrals. And, and you know, everything from GA 4 to HubSpot, there are, there are gazillion tools out there that will do your web analysis that will show you where your referral traffic is coming from. If you're if you're seeing your traffic stay steady, but your AR referrals non existent or or or persistently low, then that's another sign right of LLM still struggling to pick up in your brand and your your perspective on category and the attribute. Balancing AI Needs with a Holistic Marketing Strategy How do you ensure that you're not, I guess to kind of compare it to SEO, for example, in SEO you're not going to force keywords or you're not going to force, you know, messaging that doesn't align with your brand. How do you ensure that that's not the same case for like an LLM, that you're not just sort of catering there? Speaker 2 Yeah, I think that's a really interesting tension. And I would say that that in order to to avoid generic homogeneous language, in order to avoid an entire subcategory using the same concepts and the same language in order to describe their proposition and their, you know, and, and describe their attributes. I think it's important to see generative and agentic AI as a channel, not the only channel. It's, it's a way, it's an important additional way of driving brand visibility and reinforcing differentiation, but it isn't the only way. And increasingly, I think what's super interesting running in parallel to the, to the growth of AI is the rejuvenation of events and face to face. And you know, the events industry is boring again, you know, outside of B2B, but particularly within B2B as well, we ourselves as an agency are seeing a real appetite for face to face interaction. I think what that means is that interacting with people is is a mid funnel and lower funnel consideration. And the top of the funnel is increasingly automated. So you know, the way to avoid just using homogeneous concepts to feed AI. What it needs is to understand that yes, there is an element of AI's need facts. And therefore, in order to corroborate facts, they need the same concepts to be reiterated and echoed across the web. But beyond that, and once you've done that, finding unique ways to express your differentiation, anchoring the outcomes that you, that you serve, the needs that you exist to address, you know, your own proof points and your own operating principles and the uniqueness of your organization. And into those narratives. Once you've got that the content that you know is going to work from an LLM perspective, factoring all of that additional narrative content into and across multiple channels is a way that you can avoid only feeding the machine. And then also knowing that there is a real appetite for, you know, once visibility is established for click throughs to, you know, to, to websites for, for the information and actually an appetite for interpersonal and face to face interaction when with even more intent, when people are kind of ready and it's about, you know, broad use of the marketing mix and really truly integrated use of the marketing mix. And I really think that again, the brands that are going to win, the B2B brands and the tech brands that are going to win are those that don't, don't see their PR team as a silo from their demand Gen. team as a silo from their social, as a silo from, you know, kind of brand and marketing operations. It is not and it cannot be in lots of individual kind of fiefdom marketing fiefdoms operating within the same organization. It is one strategy, one kind of thought leadership platform or content strategy that is disseminated and brought to life across consistently across multiple channels. And I think that sense of integration is how you avoid that problem. Join Our Webinar: Navigating the New AI Marketing Landscape Yeah, absolutely. I think in our sort of complex environment, it's definitely important not to be tunnel vision there. I think that's so essential. But yeah, no, really great point there. Obviously, we do have a webinar coming up which actually kind of tackles a little bit more on like, I guess like tips and sort of how to tackle that sort of marketing environment. You know, what can we expect from that session and what's a good teaser to kind of leave things on? Speaker 2 Yeah, I think it's really important just to reiterate that point that I really don't think it's it's of any use to anyone for any age. See anyone who wants to begin to help brands navigate this to be definitive right now about anything. There are not singular right answers. There are lots of wrong answers. And, you know, I think we're all kind of finding those out in real time ourselves. But I am really super, super interested in trying every different forum and format in order to share experiences and share wins and share successes. And, and that's why I'm super grateful for this podcast and, and the webinar opportunity. And, you know, I write regularly about it from, you know, random rants on LinkedIn to, you know, to what papers. And we, we run clinics with, with clients as well, where we are, where it's simply just a, what we're frustrated about with this this week. And so from the webinar, I'm really excited about to be a little bit more of a interactive forum. And it really shouldn't be a, you know, it won't be a broadcast forum, but we will be talking about the distinction between generative and agentic. We'll be talking about in more detail about those, you know, the signals and how to audit for signals and you know, the role of four that we went through and understanding how to distribute and structure content for a genetic and generative. But also where there are opportunities to be very different and and drive down that kind of differentiation, which is always important. But there will be regular opportunities and junctures and and I don't intend to do all the talking. It would be awful if I did to just talk about how everyone that's on the webinar has is kind of factoring this into their workflow and the successes and the the obstacles that that the group have had. And I'm here to learn at, you know, as much as be of any help and I think that kind of attitude is critical as we kind of build the plane and fly the plane at the same time. So that webinar is just a bit more of an interactive opportunity. Yes, there will be things that I will expand upon that we've spoken about here and a little more. And we have got some kind of first steps that we think are useful in terms of thinking about generative and agentic. But as much as anything, it's an opportunity to share for that for everyone who turns up to, you know, to share their experience of it as well. And I'm, you know, I'm fascinated to learn so. Speaker 1 Well, I think that's the perfect teaser to leave things on. Thanks so much, Dan for sharing, sharing all your expertise. And if you'd like this conversation, we highly encourage you to register for our webinar. It's called Why Generative Energetic AI is the New Paradigm for Content Strategy. Off the back of this conversation, we will be sharing more on how to tackle this new marketing landscape as well as tips from an expert panel. It all takes place on April 23rd at 3:00 PM, so do check it out. I'll leave a link to that webinar in our podcast description. Thanks so much Dan and everyone for listening and stay tuned for another podcast. Thanks so much. Thanks so much.