In this episode
Ankur Goyal, CEO and Co-Founder of Fibr AI, explains why a brand website now has two audiences: the people who visit it and the AI agents that read it to decide how to describe the brand. He and host Greg Kihlström cover why the website is the one source a marketer controls when AI assistants can pull from Reddit or review sites instead, what an agent needs from a page that a person does not, how an adaptive website differs from segment-based personalization, and why Goyal thinks customer experience leaders, not only legal and compliance, should set the guardrails.
Key takeaways
- The website is the only source a marketer controls. Goyal says an AI assistant can ignore a brand’s site and take its answer from Reddit, G2, Gartner or customer reviews, which is why he expects the website to become the most important piece of brand information.
- Outdated third-party content can become the AI’s answer. One Fibr AI customer, a credit card issuer, has outdated card information from Reddit appearing in chatbot answers and is struggling to get it changed.
- Being cited is no longer the hard part. Being cited correctly is. Goyal says AEO and GEO work was effective at getting brands cited, and the larger challenge now is whether the agent took the right information from the site and describes the brand accurately.
- An agent reads a page without the brand context a person brings. A human visitor already has an idea of what a brand like Citi or Amex is and connects what they read to it. An agent reads the information on that page, so context that lives elsewhere on the site gets missed.
- Agents cite facts, and marketing copy often supplies none. Goyal contrasts a healthcare pitch built on a promise of faster weight loss with the version an agent needs: a specific, tested result it can cite.
- Agents miss content that is not in the HTML. Goyal says AI crawlers read a page’s HTML and miss the images, video, JavaScript and animation around it. His recommendation is to repurpose the same content in a structure an agent can consume, not to build a second site.
- The agent is one more audience intent. Fibr AI already treated a Google Search ad visitor and a Meta ad visitor as different intents. Goyal treats an AI agent as a third, served from the same website and infrastructure.
- Customer experience should set the guardrails, not only compliance and legal. A debt consolidation client ran messaging built around the keyword “bankruptcy” that cleared compliance, legal and regulatory review and matched visitor intent, and conversions went down.
- Marketing work moves toward guardrails and judgment. Goyal expects tooling and marketing operations to be largely automated, with marketers supplying gut-level insight, setting guardrails and checking outputs, and the CMO deciding what to test next.
Chapters
- 00:00 — Where AI search gets its description of your company
- 02:16 — Ankur Goyal’s background: Nestlé, Stanford and a women’s health brand
- 03:55 — What Fibr AI does: adaptive websites in place of one-size-fits-all
- 06:42 — Bots as an audience and the “agent experience”
- 07:33 — Why the website is the one source a marketer controls
- 08:51 — What it costs when a machine reconciles inconsistent product descriptions
- 10:46 — Cited versus cited correctly: where AEO and GEO stop
- 11:17 — One page, two audiences: persuasion for people, facts for agents
- 14:45 — Does designing for agents improve the human experience?
- 17:12 — Personalization versus an adaptive experience
- 21:08 — What a CDP, CMS and testing stack can’t do
- 24:00 — Who owns this: marketers as guardrail setters
- 25:14 — The “bankruptcy” example: compliant, relevant, and conversions fell
- 27:49 — One year out: will websites still exist?
- 28:47 — How Ankur Goyal stays agile
Why the website is the one source a marketer controls
Goyal’s starting point is that AI assistants are a surface marketers cannot control. An assistant can skip a brand’s website and build its answer from Reddit, G2, Gartner or reviews that may not describe the brand correctly. He gives the example of a credit card customer whose outdated card information, sourced from Reddit, now appears in chatbot answers. The website is the one place where the brand decides what the information says and how it is structured, which is why he expects it to matter more in AI search, not less.
Being cited is not the same as being cited correctly
Goyal says AEO and GEO were effective at getting brands cited in AI answers. The question he now hears from marketers is different: when the citation appears, did the agent take the right thing from the site, and does the answer describe the brand accurately? He expects that to be the more important problem going forward, and he says only the brand marketer can connect the pieces of context an agent needs.
What an agent needs from a page that a person doesn’t
A person lands on a product page already knowing something about the brand and fills in the rest. An agent reads the page in front of it. Goyal describes most agents as factual: a marketing line promising faster results persuades a person, while an agent is more likely to cite a specific, tested outcome. He points to healthcare advertising as an existing model, where a tagline already has to sit alongside a disclaimer, and describes the direction as a strong tagline with a supporting fact behind it.
The agent is a third audience intent, not a second website
Goyal does not recommend building new content for agents. Websites are a single surface, mostly served as static pages, and AI crawlers read the HTML while missing the images, video, JavaScript and animation. His approach is to repurpose the same content in a format an agent can consume. Fibr AI treats the agent as another audience intent, in the same way it distinguishes a high-intent Google Search visitor from a quick-click visitor arriving from a Meta ad.
Why a CDP, CMS and testing stack don’t adapt on their own
Goyal describes the website as the only place where paid acquisition, lifecycle, experimentation, personalization, content and SEO teams meet, and says those teams and their tools rarely talk to each other. A CDP serves a lifecycle marketer well, but it does not carry the original paid-search intent that brought the customer in. Kihlström adds that a rule could handle one path if every customer followed it, and customers do not. Goyal calls what Fibr AI adds an intelligence layer: it takes context and signals, changes the page, learns from the result and changes it again.
Who should own adaptive experience
Goyal says conversations about experience have focused on compliance, legal and regulatory sign-off, and that the customer experience leader should be the one setting guardrails, because that person understands customer emotion and brand tone. His example is a debt consolidation client: messaging around “bankruptcy” passed every review and was highly relevant, and conversions dropped because the word creates fear in the customer. He expects marketers to spend more time on inputs, guardrails and checking outputs as tooling and operations are automated.
FAQ
What is a website’s “second audience”? AI agents and bots, including search crawlers and AI assistant agents, that read a site’s content to decide what to say about a brand. Ankur Goyal treats them as an audience alongside human visitors.
Why does Ankur Goyal say the website matters more in AI search? It is the only source of brand information a marketer controls. An AI assistant can skip the site and draw on Reddit, review sites or analyst listings, which may be outdated or describe the brand incorrectly.
What is the difference between being cited and being cited correctly? AEO and GEO work has focused on getting a brand cited in AI answers. Goyal says the harder problem now is whether the agent took the right information from the site and describes the brand accurately.
What does an AI agent need from a web page that a human visitor does not? A person responds to a persuasive pitch and brings prior knowledge of the brand. An agent reads the page on its own and looks for facts it can cite, so claims need supporting facts in the page’s HTML, not only in images, video or animation.
Who should own adaptive website experiences on a marketing team? Goyal says the customer experience leader should set the guardrails, not only compliance and legal. He cites a debt consolidation client whose “bankruptcy” messaging passed every review and matched visitor intent, and conversions still went down.
What does Fibr AI do? Fibr AI builds adaptive websites: one site whose content changes based on who is visiting, including the ad or channel they came from and whether the visitor is a person or an AI agent. Goyal says the company works mainly with regulated industries such as banking and healthcare.
About AJ Goyal
Ankur “AJ” Goyal is Founder & CEO of Fibr AI. He has spent over a decade leading ventures across AI, digital marketing automation, and consumer product innovation. He specializes in personalization and marketing infrastructure, shifting web platforms toward agentic and automated consumer engagement.
Resources
Visit: Fibr AI
Companies put enormous effort into bringing people to their websites, but once someone arrives, most sites still show everyone the same generic experience. Fibr adds an intelligent layer to your existing website, without requiring a rebuild, a CMS migration, or engineering support for every update, so one URL can deliver a unique, relevant experience based on what each visitor is trying to accomplish.
And today, your visitors aren’t always people. AI search tools and agents are also reading the web to decide what information to surface. When a brand’s own content isn’t clear or useful, they may rely on sources like Reddit comments and third-party reviews instead. Fibr helps make your website a clearer, more useful source for both people and AI systems, while keeping your brand in control.
The hardest part of martech right now is telling which tools are actually earning their keep. That’s the question running through the MarTech Summit Boston agenda on October 27. We’re attending MarTech Summit Boston on October 27 as a media partner. If you want to join, the code AGILE15 saves you 15%.
Need to evaluate martech, AI, or CX platforms and don’t know where to start? The Agile Brand Guide List is now in beta. We’re building the most useful, free-to-use, and up-to-date way to assess what truly matters when evaluating platforms.
Reach your customers with Reddit. Spend $500 in ad spend, get $500 back in ad credit!
Chaser is the only Slack-native project management platform that helps teams turn messages into tracked tasks, automate follow-ups, and maintain team-wide visibility, without adopting another tool. Now integrated with Claude and other GenAI tools. Learn more at trychaser.com and use code AGILEBRAND for a 3-month free trial (normal trial is 14 days).
Start building your own apps with Replit and get $20 off.
Enjoyed the show? Tell us more and give us a rating so others can find the show.
https://www.linkedin.com/in/gregkihlstrom
Don’t miss a thing: get the latest episodes, sign up for our newsletter, and check out Martechipedia, the wiki for marketing technology.
The Agile Brand is produced by Missing Link—a Latina-owned strategy-driven, creatively fueled production co-op. From ideation to creation, they craft human connections through intelligent, engaging and informative content.
Transcript
[00:00:00] Greg Kihlström: Hi, I’m Greg Kihlström, host of The Agile Brand, and here’s a question for you. When AI search describes your company to a potential buyer, where is it getting that description and would you sign off on it? The agility question here isn’t whether you can publish faster, it’s whether the machine reading your site arrives at the same answer your best salesperson would give. Today we’re getting into what happens to the brand website when the models around it keep moving. We’re gonna cover why your owned properties become your most reliable control surface when the platforms around them are least stable, what actually changes when a single page has to satisfy a person and a machine at the same time, and where the decision about what a visitor sees has to live once it can’t live in a pre-built segment. To help me discuss this topic, I’d like to welcome Ankur Goyal, CEO and co-founder at Fiber AI. Ankur, welcome to the show.
[00:02:16] Ankur Goyal: Greg, great to be here again.
[00:02:19] Greg Kihlström: Yeah, yeah, looking forward to talking with you. I know you were on the, The One Amazing Thing show as well, and got– we got to see, see things in action, so looking forward to talking about, about all this stuff with you here today. Before we dive in, and for those that didn’t catch the u- the other episode, why don’t you give a little background on yourself and your role at Fiber AI?
[00:02:38] Ankur Goyal: So, so I’ve been– Greg, I’ve been a marketer for almost a decade now. started my career with Nestle and CPG, and I have done CPG, majorly. I, I, I went to school at Stanford, and that- that’s where I started, my entrepreneurial journey. My first brand was in women’s health, and I learned a lot there. And everything that Fiber does today is because of that brand, right? I realized that, the way we were positioning and telling about our products to our customers on Meta and on Google were great, but as soon as they were clicking them and landing on a website, they were not able to, like, see that.
[00:03:16] Ankur Goyal: And if, if, somebody’s coming from a particular, unique selling proposition, and they don’t find it on the website, they just kind of assume that you don’t do it. So-
[00:03:27] Ankur Goyal: So that’s where, like, the whole idea of, Fiber came in is why are we treating, why are we treating websites as secondary as a marketer, right? And, and, like, and then we started, like, building Fiber, like, we started three years back. The whole idea was websites are important. We need to make them important and get those surface, aligned to how ads and how other paid media, ecosystem of marketing has been doing.
[00:03:55] Greg Kihlström: Yeah. Yeah. Makes sense. And, and maybe just to dive in a little bit more, could you talk us what, what exactly is Fiber? What was the core problem that you saw in the market, and, and who are your-
[00:04:10] Ankur Goyal: So Fiber– So the core problem, that Fiber solves is basically creating adaptive websites. The idea is if you’re getting different segments and different cohorts of consumers and customers on your website, why do all of them see the same message? Why do all of them have the same experience, right?
[00:04:30] Ankur Goyal: As, consumers, we have moved ahead and we have advanced a lot because when we go to a Twitter feed, it’s, it’s very different for Greg and for Ankur, right? when we go to a LinkedIn feed, when we see Instagram, when we go and search on Google, completely different stuff, but unfortunately, when both either Greg or Ankur clicks that tweet and go to the brand website, it’s the same. And-
[00:04:54] Ankur Goyal: …it just felt too rudimentary to me, and that’s where Fiber started. We were like, “Okay, why can’t we make sure what Greg sees in my paid media or in any communication that I put out there is similar to what he sees on the website?” And that’s, that’s how marketing was, right? Like if you think twenty ten, the era of, loyal interests and, Ogilvy method, like every touchpoint should have the same messaging. It just-
[00:05:21] Ankur Goyal: …with, programmatic, marketing with Meta and Google Ads and, and TikTok and everything, it just became hard to do for a marketer. And thus, I think we somehow lost the plot and started doing it one size fits all on the website. So that’s what Fiber-
[00:05:37] Ankur Goyal: …is solving. we typically work with, right now, regulatory industries, banks, healthcare, anybody who has, who has a customer who will be, who will be with them for long, like a life– high lifetime value, is what, we focus on, as our customer segment.
[00:05:57] Greg Kihlström: Great. Great. So yeah, and so that, that brings us to where I wanna start today at least is so, there’s, there’s, there’s one thing- To personalize content to customers, but we, we wanna be in control. as a, as a brand, we want the control over the consistency, right? And so as anybody who’s been chasing model versions and, the variance between LLMs and, and all of that, it’s a, it’s a moving target, right? So, what’s actually durable here? what, what can a brand build-
[00:06:32] Greg Kihlström: …or own that still pays off regardless of which model happens to be in front of the customer, this year, let alone next year?
[00:06:42] Ankur Goyal: Sure. Greg, let me add also like, to the Fiber intro a little bit more, right? Like-
[00:06:48] Ankur Goyal: …I mentioned a lot around how humans will experience the website, but if you start thinking around what Cloudflare CEO has been saying and what people have been, realizing that almost half the traffic is now bots, right? And it’s not just the experience that humans see, it’s also the experiences the bots see, right? could be a Google SEO agent, could be a ChatGPT training agent or a Perplexity coding agent or, or anything else that’s coming on the website. They are also consuming content on your website, right? And that’s, that’s another thing that Fiber is also ensuring that what is the agent experience looks like. So for us, whether it’s an AI agent or a human or a person, all of it’s same, it’s a, it’s an audience. How do you make sure your website is adaptive to that audience is what Fiber solves.
[00:07:33] Ankur Goyal: Now, moving on to your question, Greg. I think in this era, websites will become the most important piece, of information out there because that’s the only piece of information that is controlled by a marketer, right? Like, I think ChatGPT, Clots of the world are the biggest fear for marketers because that’s– the– this is the surface which they cannot control at all.
[00:07:59] Ankur Goyal: a ChatGPT bot can choose to ignore your website, go to Reddit, or they, they let go of Reddit, but yeah, go to like a G2 or a Gartner and consume data there and put it, put it out there. Or they might consume some, some reviews out there, right? Like, which might not reference your brand correctly. Like we have a customer, who has, credit cards, and one of the credit card information comes from Reddit, and that information is outdated. And now this customer is struggling how to change that on, on a, on a chatbot, like a GPT or, or, or like a Perplexity, right? That, that is the power of website. Like right information, structured in the right format, whether it’s for human or AI, it’s secondary, but having that structured
[00:08:45] Ankur Goyal: is– will be the most important thing in marketing going forward. [gentle music]
[00:08:51] Greg Kihlström: And even, on a more subtle level, even within an enterprise, they’re often describing a product slightly differently in various places, right? a pricing page is gonna talk about things a little bit differently than support documentation or a partner listing, and we are relying on these LLMs to reconcile these things in the right way, in the right context, and, and as you can imagine, it’s, it’s not exactly what we as, as brand marketers would, would want. So what is that– what, what’s the real cost of that type of inconsistency now that a machine is, is reconciling for us to, with, to your point, very little control from the marketer?
[00:09:37] Ankur Goyal: I, I agree. I, I agree. Like I think this is a very important point because, like I was discussing with a marketer a couple of days back, and the whole point was, how would a, how would a GPT show you, right? If you think about a human visiting a website, as humans, we already know, oh, if I’m talking about, say for example, a Noom or a, a Amex or a Citi, Citibank, right?
[00:10:01] Ankur Goyal: I already know what Citibank is. I have some brand idea on it, and then I’m going to a webpage and I’m reading about a product, so I get some idea from that. So humans kind of like they’re not seeing the information that’s there, but they’re also like linking a lot of information that’s out there. When it comes to, AI, when it’s answering, it’s going to that page and just in– reading that information. Now, where is the brand context? Where is the context that exists on a product page which might be different than a pricing page? Those kind of get missed, right? And th-that’s the reason like th-that becomes a lot more important, and that’s why I, I keep talking about like agent experience. It’s very, very important to figure out what the agent can consume and how they’re going to, talk about you in future. And, and these pieces only you as a brand marketer could
[00:10:46] Ankur Goyal: connect back and give it to agent, right? You can get cited, great. But with that citation, is the citation perfect or is– did the agent scrape the right thing from your website? Those become a lot more important problems going forward, not, not what we have been talking about. And I think AEO, GEO has been great in the past, and they were great to get you cited, but now the bigger challenge is to get you cited correctly, to get– to have your brand, mentioned correctly. And I, I think that will change over time, here.
[00:11:17] Greg Kihlström: Well, and I like that you brought up the, the agent experience ’cause I, I think that we spend a lot of time on, whether it’s user experience on a specific platform or customer experience just more, more broadly. But we are– to your point, a, a webpage has gotta serve two audiences, right? So, you know-
[00:11:37] Greg Kihlström: …we’ve gotta persuade that, that human with some, some more, whether it’s aesthetics or just the way that we talk about the brand and the brand voice and, and some more emotional elements, but we’ve also got to make sure that an agent can get a binary yes, no, does this apply or at least quantify some of these things that we, we as humans kinda take for granted, right? again, like, what is– what does luxury mean to a, a bot, right? You know-
[00:12:07] Greg Kihlström: …we’ve gotta quantify what that means. So- Where do you, w-where do you see these things actually, conflict and not, not just differ? And, and how do you, how do you recommend that, that those things get resolved?
[00:12:22] Ankur Goyal: I, I, I think, like, it’s spot on, right? Like, most agents are factual. They’ll read a review. They will know. Like, uh-uh-I have a healthcare customer, and the marketing pitch would be, “We will get your… get, get you to lose your weight faster,” or s-something like that, right? And-
[00:12:38] Ankur Goyal: …but the agent pitch needs to be, “2,500 out of the 3,700 people that we tested lost 67% of the weight.” It’s a very, very different world out there, like, uh-
[00:12:48] Ankur Goyal: …I, as a human, might not decide on that, that statistics, but an agent might cite you on that sa-statistics, right? And so I think-
[00:12:56] Ankur Goyal: …I think, I think as marketers, we already know that the experience needs to different. The only thing that has to be figured out is how can you create that experience, can create different experience, right? Unfortunately, website is a single surface. Unfortunately, most, most, methodology in which websites are hosted is a static ecosystem. Uh-
[00:13:18] Ankur Goyal: …and so there’s l-very little control. if you read about how agents are consuming, how a Google Gemini bot is going or a ChatGPT bot is going to your website, they go and read the HTML, but i-in a structured format, and you can do that, that’s great, but they then misses all the images that you have put on the page. They miss the videos that’s there as a context. They might miss some JavaScript or animation that you have put in, and so that becomes kind of like a tougher. So I personally feel the goal is not to have, goal is not to, not, not to create something new, but to u-utilize, repurpose the same thing, but in a methodology, in a structure, in a format that an agent can experience, and I think that’s where, that’s where, like, agent experience
[00:14:03] Ankur Goyal: comes in. So when we used to, like one, one year back when we were just talking about human experience, we used to say, “Oh, somebody coming from Google Search ads is high intent versus a Meta ad is like a, more like a, a quick click shopper.”
[00:14:18] Ankur Goyal: agent is the third, third layer to it, right? Like, so when you are doing websites for different audience intent, the next, next audience is agent. You are doing a website experience for an audien- agent audience intent as well. So that’s how I think at Fiber, I, I think brands should handle it, and at Fiber, that’s what we are trying to do, like providing a different experience for an agent for the same website, same structure, same fundamental infra that exists for the brands.
[00:14:45] Greg Kihlström: I mean, do you think that thinking through some of this stuff from the agent experience makes us think better for the human as well?
[00:14:58] Ankur Goyal: I, I think so. Like, I, I think my bigger… Okay, so two things here. If you’re thinking through agent experience, you’re thinking through facts.
[00:15:07] Ankur Goyal: Sometimes as marketers, we go pure emotion and then put some things out, but when you’re starting thinking through like a second audience as well, I think that also, bleeds into your human experience when you’re creating it. and I think it’s vice versa, right? Like, I know agents don’t consume emotion, but I feel within n-by next year we might be talking about, hey, now agent also has some kind of emotion, and then we have to tackle not just the facts, but also the emotions to the… I think this, this whole experience layer will bleed, whether it’s into human or into agent experience. But, but I personally feel like creating, personalized or creating unique experiences for each of those segments, cohorts, agents, humans, is actually a much better marketing, uh,
[00:15:52] Ankur Goyal: methodology than creating like a single static variation for everybody.
[00:15:57] Greg Kihlström: Yeah. Yeah. I mean, I, I feel like, well, ear-early days of SEO was a little different, but like, I feel like once SEO matured, thinking through how you structure content, like I think that made a lot of content actually better. again, early days you could just keyword stuff and do all that stuff. But like, once it got m-more mature, so yeah, I f-I feel like this is similar of like, it’s, it’s kind of forcing brands to think about, well, again, how do I quantify that my car is a luxury car versus just saying, showing some fancy pictures or videos or, or something like that.
[00:16:33] Greg Kihlström: So I know it’s a little more sophisticated than that, but you know what I mean. So yeah.
[00:16:37] Ankur Goyal: No, I, I agree. Like, I think one example here is I think o-one segment that has been doing this for a while is healthcare, right? Like, you cannot put a drug out there and claim some happy customer and everything. Like, you also have to have like a disclaimer at the bottom. I think that’s like, that’s kind of like a human agent experience-
[00:16:58] Ankur Goyal: …mix, on top of my head. but I think that’s, that’s essentially where the world is going. You have like a great tagline, everything, but you also have like a supporting fact behind it so that you get cited by an agent. So that, that’s how I see it.
[00:17:12] Greg Kihlström: Yeah. Yeah. So I wanna talk a little bit more. Ag-again, I know, we got a chance to look at, at, Fiber on, on One Amazing Thing, my, my other show, but I wanna talk a little about your approach to this, to this adaptive experience. And, and so certainly personalization, I mean, we’ve been talking about personalization for, for years and years. I think it’s getting better now. I think, some of the, the, the generative stuff is actually helping with that. But I do think that there’s a big difference between something that is personalized, whether that’s basic substitutions or even a little more nuanced and truly a-an adaptive experience. Could you… what, what is your philosophy on this? how do you, how do you approach this as something that isn’t just
[00:17:57] Greg Kihlström: what a lot of people might just call personalization?
[00:18:00] Ankur Goyal: No, I, I think like personalization, experimentation, like tho-those have been words out there. The great teams have used it efficiently and have been able to do wonders with it. I personally feel overall, like as marketers, we always want to personalize, right? Like we just end up choosing the areas and, and the, and the channels where we can actually do it without depending on mo-more folks, right? So, so if you think about like a Meta or like a Google campaign or TikTok videos or, or even like organic channels, location, et cetera, we have been g-great at personalizing those stuff because these platforms provide you that opportunity to create hundreds of it. And with AI it has become a lot more automated and a lot more
[00:18:45] Ankur Goyal: faster as well. Unfortunately, all of this, when lands on your own, own branded content page, which is your website, gets lost because there an experimental personalization requires input from a tech person. Analytics is sitting with somebody else, brand or content sitting with somebody. It’s just like a complex number of people who, who are managing, this entire ecosystem. And I think that’s where Fiber comes in. Like our goal is figuring it out what you are doing on other channels, connecting that back and making sure automatically adapting the websites to it. So if I’m targeting for like podcast creators like Greg, right? Like and, and versus I’m targeting like YouTube video creators, my pitch
[00:19:30] Ankur Goyal: in ads would be very different, but my website might be similar. How do you take that information from the ad, connect it into the website and show so-something different, right? Now the, the advanced layer it is how do you figure out whether a human is visiting your website or a agent is visiting your website and change the experience for them, right? Or like if somebody’s coming from a California versus New York. That’s what at Fiber we are trying to do. Very simple philosophy. You have context and signals which could be, which could be like what the customer or prospect have consumed before. Then you know what, what is that audience or customer prospect is, which is your, their lifestyle or their demographics data, or whether they’re human or AI, whether a ChatGPT agent or Perplexity
[00:20:15] Ankur Goyal: agent and stuff like that. And you use both of these information to figure out what I should make sure change on the page so that it re-relates the best to them. make that change, run it, learn from it, and then re-change it. That’s what Fiber does. Like Fiber is a, is one area which makes sure whether it’s personalization, experimentation, agent experience, all of it is powered through the same, adaptive learning loop and every new interaction that’s happening on your website or any other channel where the customer is interacting with you gets captured and gets applied to. So that’s what at Fiber we are trying to do. it’s– You can always call it personalization, you can always call it experimentation. I personally call it its intelligence layer. We deci- we figure out based on
[00:21:00] Ankur Goyal: all the context and all the signal we have, what we want to run and what we want to showcase to our customer.
[00:21:08] Greg Kihlström: I do wonder a lot of, a lot of enterprises, they have their CDP, they have their CMS or DXP, they’ve got-
[00:21:15] Greg Kihlström: …they’ve got a testing stack. So, and certainly the intent of those things is to go beyond, basic, basic kind of substitution, first name, insert first name here kind of personalization. But I guess may-maybe just to, to dig a-
[00:21:32] Greg Kihlström: …a little deeper there, it’s like what, what are the limitations then? You talked a little bit about external context and, and, and stuff like that. what, what are the, the general limitations of some of those existing stacks that this is-
[00:21:45] Greg Kihlström: …this is building on?
[00:21:47] Ankur Goyal: I think like, it– So let me, let me zoom out and talk about the marketing stack first, right? Like if you think-
[00:21:52] Ankur Goyal: …about website, that’s the only piece in your entire marketing stack that kind of talk to external internal function. So we have a paid acquisition team in enterprises, we have a lifecycle marketing team, we have an experimentation team, we have a personalization team, we have a content writer, we have an SEO team. There’s like so many teams together. All of these teams never talk to each other. The only surface where they kind of talk to each other is website.
[00:22:17] Ankur Goyal: Now website stack is CMS and then, and the individual tools of these, all these teams. Now the problem that Fiber is trying to solve is how do you basically learn from all these surfaces on the website-
[00:22:32] Ankur Goyal: …and make sure on top of that CMS action it out for them, be it like action as an insight, action as an hypothesis, action as a personalization, action as an ex- agent experience. That’s what we’re trying to… That’s, that’s a difference, right? Like a, a CDP is great for a lifecycle marketer, but how can I use the CDP content and connect it back to an ad so it’s a pa- somebody saw a paid search, became a customer, became a CDP, now you’re sending an email to them. But what if you can also consume the paid search initial intent-
[00:23:05] Ankur Goyal: …and use that?
[00:23:05] Ankur Goyal: Like so that’s where Fiber sits and it’s, it’s basically taking all that information running on top of the website, and I keep saying website, but eventually the goal is it could be an experience anywhere, right? Like once we have that much data that will understand it could be experience anywhere, right?
[00:23:19] Ankur Goyal: So that’s, I think that’s where the limitation comes, like systems don’t talk to each other. AI helps us get it done. With Fiber we do it, we execute that, talking systems insights on websites. Eventually we might do it on like an email or go on, go up and do it on the paid acquisition. That’s how-
[00:23:37] Ankur Goyal: …I think of it, overall.
[00:23:39] Greg Kihlström: Well, yeah. And if you, technically you could probably do that, like that one use case that you described, if everyone took that exact same path-
[00:23:49] Greg Kihlström: …you could set up a rule in, in your existing MarTech side, except that not everybody does the same things in the same way and has the same intent and, and all those. So, to me that’s where-
[00:24:00] Greg Kihlström: And I think this, this brings up an important point too, which is just the, the role And the evolution of the marketer is, there, there’s a lot of rule setting and if this, then that still even as sophisticated as systems might be, there’s a lot of these like rules getting set up and, we hope the customers follow the rules that we think that we’d like them to follow, and we think they should follow. but what this ends up being in my mind is, marketers becoming about governance and okay, what, what guardrails should I put in place? What do I want a s- how do I want a system to behave ideally? That’s a big, that, that’s a, that involves marketing
[00:24:45] Greg Kihlström: ops, that involves, kind of strategy and, and thinking through multiple scenarios. How does, what, what’s your view of, who, who can own this, who should own this in a marketing team and, and what do, what should CMOs listening out there be thinking about the, the roles on their teams evolving to, to do this and do it well?
[00:25:08] Ankur Goyal: No, so, so very, very, very relevant, right? Like, uh-
[00:25:12] Greg Kihlström: Sorry that was like five questions, but yeah. [laughs]
[00:25:14] Ankur Goyal: No, that, that’s fine. Like, I, I, [laughs] I hear you and I, I was, I was smiling because I recently, got a code request on almost similar stuff. So, the request was more around in, in the, in experience world, we have been only talking about compliance, legal, regulatory. But if you think about it operationally, it needs to be the CX officer or like a CX guy, because eventually that’s the person who, who sets the guardrails.
[00:25:45] Ankur Goyal: That’s the person who knows the nerve of the ma- nerve of the customer, like, a emotion or a brand tone. Like for example, we have a customer where we talked about, bankruptcy, right? Like they are a debt consolidation player. We talked about bankruptcy as a keyword. We did a great matching, everything works personalized, high- highly relevant and, and the, and conversion went down. Now this was completely aligned to a compliance officer, a regulatory and a legal guy. But the conversion went down because it was not aligned to a customer experience, folks-
[00:26:19] Ankur Goyal: …because they know that bankruptcy creates a sense of fear in my mind as a consumer, and that not, might not be the right way to pitch it. You might pitch as-
[00:26:31] Ankur Goyal: …debt consolidation or solving for your debt, et cetera, but not bankruptcy because just, it basically you are jumping so many hoops that he’s already in the mindset of bankruptcy, right? Like I think customer experience overall and overall marketing will evolve to more guardrail setting, more understanding the insights, adding the gut, gut insights. I don’t think AI can still do the, the, gut level insights, adding those insights into the system, checking outputs, figuring those outputs are right or wrong, like tweaking them, giving kind of emotion that you have to. Like, that’s the guardrail layer a marketer will work. Tooling will become mostly automatic, right? Like, like what Fiber does. There’ll be some other tools who will automate it with AI, et cetera. Eventually the input signals and the output is where
[00:27:16] Ankur Goyal: the marketers will play. The ops, markups piece will get completely automated, and I think that’s where I, I would see the marketing organization change a lot. Like a CMO will be the more orchestrator like they are right now, but-
[00:27:29] Ankur Goyal: …more orchestration on what to test next, what new emotion to try and et cetera. and like the teams below are more like guardrails on, okay, if we are testing this, what kind of guardrails we should put in and how would the, h- how would, how should the output look like and how long we’ll do it. I think that’s where the, the marketing org will move to.
[00:27:49] Greg Kihlström: Yeah. Yeah. Well, Ankur, I could probably talk about this all day, but [laughs] thanks for, thanks for joining today. couple last questions for you as, as we wrap up here. first one is, if we were having this interview one year from today, what is one thing that we would definitely be talking about?
[00:28:07] Ankur Goyal: [laughs] so one thing last year also I, I said the same thing f- to somebody is after one year we might be talking about website should not exist or we might be talking about website is the only way to exist, right?
[00:28:19] Ankur Goyal: [laughs] That, that’s where I think we will be because like with AI, innovation, I don’t know how the internet will eventually look like. Will it be, will it be, will it have the fate of the yellow, yellow pages that we used to have, and everything is like chat based and voice based and nobody goes, [laughs] goes to that? Or will it have a fate like, oh, this is the space that I have to optimize on, right? Like, so that’s one, that’s, that, that’s my answer to that question.
[00:28:47] Greg Kihlström: Yeah. Yeah. And last question for you, what do you do to stay agile in your role, and how do you find a way to do it consistently?
[00:28:55] Ankur Goyal: I think, for me, two things that I do to remain agile in, in overall space and what, what I do, in my role. One is I do significant context switching. So being some, s- like s- sometime I am in product, then I’m like talking to the tech team, talking to the customer, being on the sales call. I think those, all of that gives me a significant, exposure to things-
[00:29:20] Ankur Goyal: …that I can think and wire all of them together and, and, and like be aggressive and be agile, overall. And I think, the second, second one for me is, more about, how do you context switch but also be closer to the customer.
[00:29:37] Ankur Goyal: that’s how, that’s how I typically do it. Consistency is something that’s like, like if you, if you… Like it’s kind of like my second startup, so I, I’ve been like used to it. So I’ve been able to do it again and again. You always have your low days, but most days I think I have been consistently able to do it just to make sure that I, I love to learn more. I love to like, weave threads and figure out insights and et cetera. So it, it, it has been consistent for me that way.
