Earn Trust In AI Discovery with Search, Social, and Media

Catch the key takeaways

As AI becomes part of early discovery, consumers are asking longer, more detailed questions and moving across search, social, and cited sources to validate what they find. Our recent webinar explored how brands can respond by focusing on the audience decisions that matter, building credible proof, and coordinating channels around the same consumer need.

Here are four takeaways to help your brand prioritize AI visibility, reinforce its claims, and measure whether that work contributes to business outcomes.

1. AI is capturing more context earlier in the journey

Consumers haven’t developed entirely new needs because AI exists. They’re expressing those needs with more context. A search for “best checking account” might become a detailed question about low balances, recurring fees, and what to consider when switching banks.

That added detail reveals why someone is looking, what’s frustrating them, and which factors may influence their decision. Consumers may begin that process in an LLM, then move to search, social, or cited sources to validate the answer. Brands need to understand the full question and the places consumers visit next.

2. Prioritize questions by audience and business value

AI visibility tools can generate extensive prompt lists, but volume alone doesn’t determine which conversations deserve attention. A useful question reveals something about the audience, where they are in their decision journey, and the potential value of reaching them.

Start with the questions closest to conversion, then evaluate the effort required to appear for each one. Paid search and media data can help confirm demand, reach costs, conversion behavior, and audience quality before a brand invests in closing a visibility gap.

3. Brand claims need credible outside validation

Useful content answers the consumer’s question and helps them take the next step. Credibility develops when authoritative outside sources support the same claim. LLMs may cross-check brand content against publisher coverage, Reddit conversations, reviews, social platforms, government websites, and other sources.

That makes consistency across owned, earned, and social content important. A brand can state its positioning, but that claim remains a hypothesis until consumers, creators, and other trusted voices reflect it back through genuine experiences.

4. Every channel needs a role in the same audience decision

Cross-channel activation starts with a shared audience insight. Paid media can test which messages lead to meaningful engagement, while social listening, search behavior, creator content, and AI visibility reveal how that message appears across different touchpoints.

Measurement should reflect the job assigned to each channel. That may include brand presence in AI answers, citations, source visibility, message adoption, qualified website actions, incremental lift, or business outcomes. The process continues as teams refine the question, strengthen the supporting evidence, and apply what they learn to the next audience decision.

FAQs

Where should a brand start when improving AI visibility?

Begin with the audience and the questions that matter most to their decisions. Evaluate each question based on journey stage, business value, and the effort required to appear in the answer. Questions closest to conversion can provide a practical starting point.

Then establish a baseline. Check whether the brand appears, how it’s represented, which sources the LLM cites, and where meaningful gaps exist. Search demand, media performance, audience quality, and conversion data can help determine which gaps deserve investment.

How can a brand become a credible answer in AI discovery?

Create content that fully answers the consumer’s question and helps them understand what to do next. LLMs are built to support a decision or action, so useful content needs enough context to guide that process.

Credibility also depends on what authoritative outside sources say. Publisher coverage, reviews, social conversations, creators, and other trusted voices can reinforce the brand’s claims. When consumers reflect the same message through genuine experiences, LLMs have stronger evidence connecting the brand to the answer.

How should channels work together around AI discovery?

Give each channel a clear role in the same audience decision. Social listening can reveal how consumers describe a problem, paid media can test which messages lead to meaningful action, and search and AI visibility data can show which questions and sources are shaping discovery.

Teams should bring those findings together and reinforce the same audience insight across content, PR, creators, search, social, and paid media. That continuity helps consumers encounter a consistent story when they leave an LLM to validate an answer elsewhere.

How should marketers measure whether an AI visibility strategy is working?

Start with the job assigned to each channel. Organic measurement may include brand presence, citations, source visibility, how the brand is represented, and whether it remains part of the answer as a consumer refines the conversation.

Social signals may include saves, shares, comment quality, and whether audiences repeat the brand’s intended message. Paid media can measure reach, engagement, qualified actions, and incremental business results. Together, those signals show whether the strategy is gaining traction and where the next round of refinement should focus.

Dive Into the Transcript

Josh Squires (00:00):

Welcome to Earn Trust in AI Discovery with Search, Social and Media. I’m Josh Squires, Director of AI Search at AMSIVE. Thank you for joining us. Today, people are using AI work to work through their problems, compare options, build shortlists before they know which brand they’re searching for. But AI doesn’t necessarily mean users trust AI when it comes to important decisions.

They’re still vetting AI answers. We’re going to unpack what’s changing about the discovery phase and how your brand can earn trust with potential customers while building proof for LLM systems. At the end, we’ll share the webinar recording more via email. Please ask any questions you have in the chat. We’ll have a Q&A at the end. After that, we’ll meet the rest of our speakers. Inna, would you care to kick us off?

Inna Zeyger (00:54):

Sure thing. Hi everyone. I’m Inna Zeyger. I’m the VP of Digital Media here at Amsive. So I oversee our paid media practice across paid search, paid social programmatic, and really anything where we’re putting a media dollar behind reaching the audience, including some of these newer AI environments. I look at things through the lens of audience, investment and measurement, so that’s really the perspective I’ll be bringing today.

Chalice Jones (01:17):

Cool. And hello, my name is Chalice. I am one of our leads of our social influencer practice, director of social influencer, and I am excited to be here today, bringing the lens of social strategy, the community, social intelligence, those kind of third party external sources and signals that end up informing and supporting what’s happening in AI.

Josh Squires (01:46):

Thank you both for joining me. First, let’s set the stage and talk about how consumer behavior is changing. On the organic side, we see people asking AI longer, more complex questions, and it’s happening earlier in the journey than we typically saw with search. The implication is that they’re giving AI a description of the information they want. With all the parameters, price, size, feature, location, on the assumption that AI was just going to do all the work and all the filtering for them and just come back with the correct answer. But at the end, they’re not always trusting LLM responses.

We’ve had studies show that they bounce around, they’ll leave the LLM and go back to Google search or Bing or social or whatever they were getting their information from. So they check it either via search or these other cited sources in the response itself.

In some cases, people will click through those cited sources, maybe not as often as we would like, but they are checking the facts. So what I would love to know, Chalice, how are you seeing user behavior change in social media?

Chalice Jones (02:50):

Yeah, so I think this is a little bit less of a change, but something we know about social is that it’s been this validation layer. I think one of my favorite things about what’s happened as AI has been becoming pervasive in the way and really infiltrating how we approach things as marketing teams is that it’s validating something that we already knew was happening, but it’s these kind of machines and larger organizations putting that final. Yeah, I guess validation on it. Let’s use that one more time. So anyways, so really what it is, it’s that research. It’s people looking to say, “Hey, is this brand really living up to what it is? I got this answer maybe in chat, but then really what’s going on with this brand?” It’s where brands are able to really present themselves, not only showcase to consumers, it’s where the rubber meets the road, make sure that they’re demonstrating it, but also where we’re able to hear back on if people are experiencing the brand in the way that we are trying to portray the brand and living up to our promises as brands.

And so it’s really this pressure test moment to make sure it’s not just fluff or whether it’s AI or brands speaking things that are maybe unsure, it’s where people come back to really validate what they’re perceiving about the brand.

Josh Squires (04:22):

Inna, what are you seeing? How’s that the same or different?

Inna Zeyger (04:27):

It’s fun to say the least, but really building on what Chalice has been talking about, this is really important from a paid search perspective because what we’ve already been seeing is that change up in how it shows up in search data. So people haven’t developed entirely new needs because ChatGPT or AI exists. What’s changing is how much context they’re giving us when they express those needs. So think about something like choosing a checking account.

So in Google, somebody might search best checking account. We know that’s a very common kind of keyword or query for a banking advertiser, but it doesn’t really tell us much about the person. Now compare that to the kind of natural language questions we’re increasingly seeing in AI and longer tail search. So it might be, “I don’t keep a lot of money in checking and I keep on getting hit with fees.

What should I look for if I switch banks?” So same basic product need, but the second question tells me why they’re in market, what they’re unhappy with and what may actually matter in making that decision. And we’ve seen over the years Google adapt pretty aggressively around this exact change in behavior.

So AI Max, we know we talk quite a bit about that, is becoming a lot more prominent. Broad match has become a much bigger part of discovery, and the platforms are really increasingly using context and signals beyond just the keywords that you’ve selected to identify that relevant intent. So AI isn’t really eliminating that search intent, it’s just giving people a much richer way to express it. And that matters a lot because paid media strategy has to account for the fact that the intent may start in one place and convert somewhere completely different.

Josh Squires (06:12):

So we’re getting more context than we had before, but the user journey’s now more fragmented than it was, their platform hopping, going back and forth. I think one of the things we saw as a challenge for brands early on was figuring out how to adapt to that change. Inna, can you tell me a little bit about how marketers can shift their thinking and adapt how they operate so they can stay ahead of that curve?

Inna Zeyger (06:38):

This is the one where everybody has to make better friends with each other, or to put it a little bit more bluntly, the biggest shift here exposes a real challenge for a lot of brands. So a brand may have a really strong SEO team, they may have a super sophisticated paid search program, maybe they own organic social internally, PR sits somewhere else, and maybe they have multiple agencies involved. And every one of those teams can be doing their individual job really, really well. But here’s the reality. ChatGPT and other LLMs don’t care about your organizational structure. So an example would be if someone asks, and I’m going to keep the checking thing going, what’s a good checking account if I’m tired of paying overdraft fees? That answer can be influenced by the brand’s own content, publisher coverage, Reddit conversation reviews, and so, so many other signals.

So meanwhile, your paid team may already be sitting on really, really valuable information about that exact audience. They’ve probably tested different messages across millions of impressions and thousands of site visits. They may already know which audiences respond to which benefits, which messages actually drive some of that more qualified action. But here’s the reality of it is that that information shouldn’t live inside a paid media report, and that’s where that shift has to happen.

So the question isn’t what’s my AEO strategy and then separately, should I buy ChatGPT ads? The question really becomes is what is this audience actually trying to decide? What information is shaping that decision? What evidence do we already have about what matters to them? And that’s where paid, organic, social and other channels can really influence that decision. And that’s a lot more useful of a conversation than simply asking, do we need to be buying another channel?

Josh Squires (08:32):

Yeah. Yeah, definitely.

Chalice Jones (08:35):

Cool. And I’ll just jump on that if that’s okay. I just want to jump on that a little bit because I think one of the things that Inna just mentioned both before and now is the usefulness of the questions that we’re understanding. And so that’s one of the ways where when we look outside of just the individual channels and when organizations can sync up across their disciplines, they’re able to understand more about the consumer, but it always comes back to the consumer at the end of the day.

So understanding the audience question first, so understanding what they’re trying to know, but not only that, trying to understand how they’re asking it. Inna mentioned this earlier, that natural language. One of the things that social and social intelligence, forums, these other places, we’re getting this kind of raw perspective and insight into how people are asking things, what their pain points are, how they’re even phrasing the question.

I think as marketers, we have a little bit of a tendency to polish the edges of the questions that consumers might be asking. But when we look at some of these other channels, we’re able to get that, again, more unfiltered lens of what it is that they’re saying and asking, and then we’re able to meet them where they’re at with that. And so considering how they’re saying something, understanding that that’s probably how they’re also prompting things in these LLMs, when they’re seeing AIOs now, we’re getting much more comfortable asking more informal questions, but then we’re also able to understand what the intent, as Inna said, the intent of their question is more than just with queries, but what’s the backside of that?

And so it’s really this shift from this brand first messaging of this is what we want to say to the consumer, but understanding what it is that they really are trying to communicate or try to understand what it is that they’re trying an answer to and answering their question.

I think it’s keeping an open mind and perspective to continuously want to learn about the consumer, and then not just providing answers, but being the answer to that question, focusing in on what they want to know.

Josh Squires (10:53):

Excellent point. And I think exactly like you said, the audience has to be at the center. It’s really about their questions, not our marketing message. We need to be finding where those things intersect. For any organic folks on the call, I think our role sort of becomes facilitator where it’s our job to understand how the LLM works, where and how it’s surfacing information, and what type of information it’s surfacing. And we can carry that to the other channels and say, “Look, we found these insights.”

When people ask this type of question, the LLM is surfacing these sources, these types of sources, earned versus owned. And if it’s earned, take it to your PR team, take it to corporate comms and say, Hey, here’s what we’re seeing. How can we craft new messaging or get onto these other platforms so that our messaging is more visible in the response?

But we have to keep it within the guidelines of what is going to be relevant, useful, practical to the user in terms of answering their question. And I think that’s something that, not to self-brag, but I’m going to self-brag, that’s something that’s at the center of Amsive’s marketing has been from the very beginning. We’re very centered on our audience science approach, which allows us to build strategies around the audience. And because we’ve been doing that for a while, this is less of a major turn for us than maybe it is for some folks.

We’re basically taking really deep knowledge about an audience, breaking it down by channel, and then figuring out how to reassemble the strategy. I think to help break it down for folks on the call, we’ve sort of assembled it into a framework that we’re going to share. We’re going to talk you guys through it a little bit.

Shorthand, we’re going to call it CLEAR, and it’s obviously an acronym because why wouldn’t it be? We’re in marketing. So once you understand the audience and you can better assess your visibility in LLMs, you can use this framework like the one we’re going to discuss here, and just know that you’re not going to be able to be strong everywhere. We’re not going to come out and own page one the way we used to on search 20 years ago. This is going to be about really honing in on your audience and their key questions and starting with what matters the most, the highest value to the business, and then fanning out from there.

So let’s get into this framework. Chalice, how do you choose which audience questions and conversations are even worth focusing on? How do you start that process?

Chalice Jones (13:27):

Yeah, that’s such a good question because I think I was just saying this, there’s so much information out there in so many places we can gather that, but then how do you focus in? And I think that as we are able to get more, what’s the best way to say this? As we’re able to get more specific with the audience, we can kind of move out of this approach of being everything to everyone. Sometimes when we’re everything to everyone, we’re nothing to nobody, which is probably not proper grammar.

But I think what’s important about that is that we’re able to actually get into the nitty-gritty about how we intersect with what our customers’ needs are. Like I said before, it’s not just answering a question, it’s being the answer to their question. And so when we understand them a little bit more deeply, we’re able to demonstrate that we understand them.

We’re not just answering it, we’re showing them that we understand their pain point. I think on the organic social side, there’s a lot of, I mean obviously in marketing we talk a lot about human behavior and psychology and all those things, but what I love selfishly about organic social is that emotional and that resonant, that human connection element of it. And I think consumers continue to get, are very smart. They know how to sniff through the fluff that might be coming at them, and that’s why brands that make them feel seen, understood, and heard ultimately, which is why I joked about being a little relationship-y, but those individuals or those brands really stand out because that is the intersection of where the need and the service or the product intersect, and that’s what makes it so powerful. And so those signals end up giving us information about how we’re able to understand how they’re asking questions and then show up in the ways.

And that just furthers our credibility on how we can solve it because we’re understanding who they are. The last thing I’ll add to that too when it comes to audience is, I kind of mentioned this before, but staying curious about the audience. If there’s something we know is that they’re going to be changing, and unless you have a product that’s for one person and that goes with their life and then your brand ends at the end of that, they’re always going to be changing.

I know I’ve mentioned this to a lot of our health clients or anybody in Medicare, but what it is to be 65 today is not what it was to be 65 10 years ago. So we have to constantly be learning about what their behaviors are, where they’re validating, where they’re learning new information, what those inputs are, and adjusting accordingly and letting that be how we choose where to meet them and what we’re going to do.

Inna Zeyger (16:18):

And I want to just quickly build on that. To your point, one of the biggest risks with AI visibility right now is that we can generate enormous prompt lists and convince ourselves that we need to show up for all of them. You said it way better than I did at Chalice, but we don’t. So maybe going through an example. So take small business insurance, it’s a category, but that compared to I just started a landscaping company, I have five employees and two trucks, what insurance am I legally required to have? That’s an actual business problem, and that’s where that distinction really matters. So the person who’s searching business insurance quotes may be ready for paid search right now.

The landscaping owner trying to figure out whether their personal auto policy covers the truck they’re using for work is way, way earlier in the journey. So that’s where there may be a much better opportunity for paid social or video programmatic or even an AI environment where we can introduce the brand before they’re ready to request a quote.

So I don’t want to choose conversations based on just the prompt volume alone. I want to know what the question tells me about the audience, where they are in their decision journey, and whether there’s a meaningful business value there. So really the way to think about it is the audience and business value comes first, the keyword prompt and platform come after that.

Josh Squires (17:43):

Yeah, I think you guys are really same brain. And the nice thing is it’s similar across all channels. We do actually all want the same things. We’re driving towards the same goal, and there’s a really similar approach. So agreeing on the questions to answer and who the audience is really important. And on the organic side, we found some value in establishing hierarchies, building out matrices of audience segments, journey stage, and value prop per each of these individual items.

So I’m going to borrow your insurance example for a moment, and bear with me for this being not totally accurate, this is top of the head, but if your best customer is somebody who’s starting a business and you’re an insurance company, the question you posed is a higher value question to Chase. We want to show up for that. We want to be there for that customer.

We want that business. Somebody who owns a business that operates in 16 counties and they’re looking to scale up to a fleet and need insurance for their new fleet, if that’s not our best customer, that’s less important. Maybe that’s a growth opportunity a little later down the road, but when you compare the time and effort it’s going to take to show up for the individual questions, it’s really important to prioritize them by business value, the effort it’s going to take to show up for the answer, and where that user is in the journey stage. Much like regular marketing and especially content marketing, we want to focus on the people who are the closest to a conversion first. We want to get the money flowing. We don’t want to build a giant audience and then have nowhere to send them or be unable to convert them at the end.

So starting from the closest point of conversion, the last couple of questions somebody’s asking to vet brands or pick the route they’re going to go, we want to be there and show up for that. So once you know the audience and the questions matter, how do you understand where your brand is showing up and where the gaps exist?

Again, on the organic side, those priority questions, you should know those off the top of your head, but there’s some more technical work, your organic teams, AI SEO, SEOs, GEOs, whichever acronym you guys are going with, they have access to mine those questions out and look for the different ways that people are saying them. As long as we’re focusing on the attributes being used, the way people are describing the thing they’re looking for, we’re not going to venture too far off base.

So we’re going to capture those questions, build that list. We’re going to get into our tracking tools and we’re going to look for do we show up? How are we being represented? What sources are being pulled for these answers and are we present on those sources? That’s generally the starting point from the organic side. Chalice, from the social media perspective, what are you guys looking at?

Chalice Jones (20:40):

Yeah, so I think similarly on the organic social side of things, a similar approach, you’ve got to meet them where they’re at, and it’s a slow burn sometimes, but that’s where all these things working together is really important. And so really it’s about mapping, I mentioned this before, it’s mapping the external story to what the priority question is for that consumer, but then looking at those additional external sources, we kind of talked about PR or there’s influencer, there’s creators, there’s all these other sources that are either validating or maybe, what’s the word, contradicting, that’s the word, there it is, contradicting what it is that we’re trying to understand or say, and that’s where we’re able to identify what are those quick wins where we can come in and answer potential gaps, and then where are things where we need to learn and maybe not necessarily set the record straight, but maybe or look at why is that a friction point there?

I like to think about it too, especially in this AI search space, we got to understand that these things are pulling in all these inputs from so many different locations, so making sure that there is some synchronicity across what all those things are saying and understanding that just the brand saying what they think about themselves is not enough. So what are those third party sources, influences, and communities that have something going on? So it’s not just, are we visible, but it’s what story does the audience encounter when they validate us at each step of the process?

Josh Squires (22:33):

What are you seeing? I feel like maybe you’ve got access to some unique data. We have information,

Inna Zeyger (22:39):

Something like that, yeah. And building upon that, because I think both Josh and Chalice, you talked a lot about what’s happening within the social space within organic, and then their paid media is, there’s so much information there, but that can help make that gap analysis just a lot more accessible. So if you’re not showing up around an AI conversation that you think matters before you automatically decide that you have an AEO problem, we can really look at what you already know.

So for example, do people actually search around that need? So that’s search queries, how much demand is there, and that’s around search volume, what does it cost to reach them? Did they convert? And then other tactics like AI max, broad match, PMax, what are those surfacing that you haven’t explicitly planned around, whether that’s content or keyword coverage or awareness around that? So then you can layer in paid social, what have you learned from paid social?

Maybe the data tells you that this is a high value audience and you’re genuinely underrepresented in AI and that’s a real gap, or you could find that maybe every paid test around that audience produces low quality traffic and no real resulting business outcomes, and that should really help change how much you care about fixing that particular visibility gap. So AI visibility can tell you where there may be a problem. The paid behavior side helps tell you whether that problem is important enough to invest against.

Josh Squires (24:16):

Love that. So just to recap our answers here, we’re talking about establishing a baseline of where the brand is present, how it’s represented, and which gaps matter for the audience you choose using internal data we already have access to vet the problem, work through the best opportunities to close them, and also validate that the LLM is the source that we should be correcting versus perhaps something going on in channel. Once you know where the gaps are, the next question is whether you’re able to close them.

So that’s going to move us on to the next question. What does a brand need to demonstrate to be credible and useful in the conversations it wants to participate in? I’m going to underline useful because LLMs really love, what’s the word? They want an action, right? The response isn’t just yes or no, a flat fact delivered to you.

When you say you need insurance coverage for your lawn care fleet, it’s going to go and do so much research, right? It’s going to try to flesh out that whole decision tree for you and bring back these answers, and that requires a lot of context and there should be some action. The LLM expects that you need this information because you need to take an expected action. And so answers need to have high utility. It can’t just be, here’s the fact, here’s the fact, and here’s the step you take next.

On the organic side, that’s kind of easy to control. We can manage that with content. In content posts, on product pages, on feature pages, whatever it might be, we can provide the information that answers the question and then step in and say, “And here’s what you’re going to want to do next with that.” And that can be your standard CTA, that can be pushing to another source, another landing page, give us your email, whatever it might be, pretty straightforward.

But then there’s a little bit of a hitch. The LLM really wants to validate the answer, and it’s going to take the answer that it retrieves from all these different websites and then crosscheck that. So that’s where credible comes in. So now that we’re useful, trying to figure out how to be credible, what sources is it going to validate your answer on your website? You said this thing and you said it like it was true. How is this machine going to know that you’re correct? Well, it’s going to go to other places that it knows those conversations are happening.

It’s going to go to places like Reddit, it’s going to go to Facebook and TikTok, it’s going to go to government websites, it’s going all of these different places. And on the organic side, we can use our visibility tools to observe what sources get pulled in, but based on what sources are pulled in, we should be taking that information to our other channels and saying, “Hey, earned media makes up 70% of the sources that are being checked for this.

How well are we covered? In all of our PR placements over the last three months, did we address this problem? Did we provide this next step? Was this phrase used in all of our earned media? What are the chances the LLM has seen and absorbed this?” That I think is our biggest role. And again, it sort of goes back to that facilitation of we’re the nexus, all the information’s passing through the LLM. The organic teams are here to catch that information, filter it, decipher it, and then take it to the people who can action on it. So Chalice, what do you look for? What do you action on?

Chalice Jones (27:55):

Yeah, so I think something that’s really important to call out is that, well, we have this lovely acronym, establishing proof, the proof, it’s not just about having outside voices, it’s having authoritative outside voices. Just because it’s outside doesn’t make it automatically credible. It’s got to actually mean something or come from somebody that’s trustworthy. Even on the social side, I tend to always throw out our E-E-A-T acronym because I think it’s really important because at the end of the day, it’s really asking creators, not just creators in the form of influencers and UGC creators and things like that, but anybody that’s developing content as a brand, as anybody, to ask more of their content, to hold it to a higher standard, to make sure that it’s not just a hack or a growth tactic to be able to get things out there. And I think that’s something that we’re going to keep seeing.

We’ve come such a long way in the last five, six years with AI, but we also have so much more to go and we’re going to keep seeing what is informing all of this change. But I think what I love about things like the EAT principles and this idea of establishing proof is that foundationally it’s about qualifying and it’s about strengthening why it is that you are the best option, the best solution, the best partner, the best X, Y, or Z.

And so thinking about it that way, it is working with external voices, making sure that they’re also holding that level of credibility. You mentioned a second ago, Josh, about it’s pulling from Reddit and Facebook and TikTok and YouTube, and that’s going to keep being true, but changing in varying forms and how much it’s considering those things. We saw a lot of that, there’s kind of a stack going around from read it from last week and all of that.

And so I think that shouldn’t scare anybody. I think it should really honestly embolden individuals to say, okay, let’s really look at how we are coming across in these authentic places and making sure that we’re doing right by our customers, meeting them where they are in these spaces, understanding where they’re asking questions and not trying to hack the system, not try to do anything, no keyword stuffing, none of that.

I know that there’s easy ways to move the needle, but at the end of the day, it’s still going to come back to is it quality? Is it answering the question? Is it providing utility? Is it providing authority in that space? So I think that’s when we’ll be able to pressure test it ourselves because we can put out there our positioning on how we want, but it’s really just a hypothesis on how we think we’re being seen until we see that kind of validated back from our audience, whether we’re seeing them actually reflect back that that statement we’re making about ourselves is true.

Josh Squires (31:04):

I think that last part’s super important. We need to not just put things out there, we need to have it reciprocated. And I think that’s really what the LLM is looking for when we talk about validation and consensus. What did you say about yourself and are people reflecting it back to you in a genuine way in organic spaces? Yeah, that’s absolutely critical.

So just to recap that section, collectively we’re building proof that connects our brand to the LLM answer and then reinforcing it through both expertise, credible outside voices. And at this point it can be really easy for in-house teams in particular to fall back into siloed channel operation. I’ve got my little bit of information, I’ve got my marching orders, I’m just going to go build things and make things and up the spend and I don’t need to go back to the well until it’s time for more data.

That’s not the best way to do it. You certainly could, but we’ve found that really working cross-channel is the way to get this done. So once you know the audience and you have the proof, how do you activate across channels without going back to siloed tactics?

Inna Zeyger (32:20):

And I think that’s a big question that everybody wants to get underneath, but this is where I think we can get really practical with that application. So I’ll use a real example from my own life because apparently I’ve spent way too much time recently asking ChatGPT about protein bars. Don’t ask me any follow-ups on that other than what we’re going to go through.

But let’s say what I was asking was what’s a good high protein bar that actually tastes good? And there’s a bunch of things that a brand could potentially tell me. So maybe it has 20 grams of protein, maybe it has low sugar, maybe it actually tastes good. And these are all really legitimate proof points, but we shouldn’t make assumptions that they carry equal weight with the person asking that question. So that’s where paid can come in and help us test that really quickly.

So I can take essentially the same audience in paid social and test creative around protein content versus low sugar content versus taste. And I don’t want to just know which one is going to get the highest click-through rate. I want to know which message actually gets somebody to engage with the product, look for where to buy it, add it to cart, and then ultimately make that purchase. But what I want to do next is I want to connect that with information that already exists across a brand’s organization. So for example, your social team, if your protein brand is probably doing social listening and seeing people repeatedly talk about taste in creator comments and Reddit threads.

Your SEO or content team is seeing longer tail searches around high protein snacks that actually taste good. Your brand agency may believe simple ingredients are the primary differentiator. Meanwhile, your paid team is seeing taste focused creative consistently outperform nutrition-like creative.

And so when you look at AI discovery, taste comparisons and product reviews are also disproportionately shaping the answers. So those are not five unrelated channel insights that I’m talking about. That’s the audience telling you something pretty consistently across five different places. So activation shouldn’t mean everybody takes that insight back to their own team and builds another separate plan. It should be, okay, we think taste is an important part of the decision for this audience.

How do we reinforce that across the places that actually matter? So what does that actually look like? Your content strategy changes to address it, so you produce more content to support what you’re seeing across those five different places. You invest more in creators actually trying the product because a creator saying it tastes good is much more credible than the brand saying it tastes good. Paid media can then go ahead and amplify that creator content.

You’ll start seeing within search, it’s picking up the longer tail language you’re seeing. You can use programmatic or video to reinforce it to the right audience. And then you can look at ChatGPT ads giving you another opportunity to really participate while somebody’s actually working through that exact question. But the key here is that I don’t need every single one of those channels simply because they’re available. I’m trying to give each channel a job around the same audience decision. And I know we talked a little bit earlier about it, but that where paid can help surface those gaps very quickly. So an example would be if it tastes good messaging gets me interested in the protein bar and get to the product page and there’s nothing supporting that claim, that’s a problem. Or if your paid data says taste is what gets people engaged and interested, but every creator review says the product tastes terrible, I cannot media plan my way out of that.

So paid isn’t creating the proof, it’s helping us understand which proof matters enough to move somebody. And really the key takeaway is that paid can surface those gaps quickly, but the rest of the organization has to be able to support the promise of what you’re putting out into the market, whether it’s a protein bar or not.

Josh Squires (36:18):

Chalice, what do you think?

Inna Zeyger (36:20):

It’s a lot of words.

Chalice Jones (36:21):

Like ditto, same, retweet. No, I think you hit the nail on the head with that, Inna. I think truly it is, we have all these inputs and I think that’s where well orchestrated marketing teams have this superpower is because you’re able to have these quick tests and learns that are kind of flowing back and forth between what you’re hearing the audience say and then what you’re able to test. We obviously in organic social, you have very loose A/B testing you can do. You’re never going to have a fully controlled environment, but then you can put the insights from that into paid and say, okay, oh, actually when we actually get a little more focused in, this is the final winner.

So organizations have the opportunity to sync that up to really understand what’s really resonating. And again, bringing that back to building that trust and authority with the LLMs, it is about making sure to what Inna was saying, that those proof points are kind of happening and being.

There was a way you just said it that was really nice, Inna, but basically putting forward the proof of bolstering what’s happening in paid. They are actually living up to that messaging or that promise that’s going out there from paid. It’s got to support those claims. So yeah, I think honestly, Inna crushed it. Yay, thanks.

Josh Squires (37:57):

I’m going to be very controversial and agree with you both.

Inna Zeyger (38:00):

Wow, how dare you.

Josh Squires (38:05):

I think again, organic’s role here is facilitator. You guys have all of the data, you’re telling the stories. What LLM optimizers can do for people working in that data set, we can still ask for, “Hey, what do you see as being the most valuable conversations in Reddit threads, in review feedback, things like that. We can make sure that onsite PDPs have the right attributes that align with our audience.” And I’m going to go back to what I mentioned closer to the start around the audience messaging matrix, knowing what the right message is for each individual audience. Inna, you had the example of the bad tasting protein bar.

Inna Zeyger (38:54):

I’ve been traumatized in the past, yes.

Josh Squires (38:57):

Haven’t we all? But there’s an audience for it. There’s absolutely an audience out there that, “Hey, that’s my 15 to 25 grams of protein today in a single bar form that is portable and clean and doesn’t require a blender, so I’m just going to bite the bullet and eat it because the protein intake is more important than the taste.” Great. So maybe there’s just some alignment issues that we need to work through first and then we shift accordingly and we go back and we re-vet our sources.

Okay, for the new message audience combination, what sources get pulled in? What do we leverage here? What can the LLM know about this person? And that’s another area we’re not going to dig into too deeply today, but personalization and LLMs are really going to be a helping hand here. It’s going to have some history about what have you talked about? Is this the fourth time you’ve asked me about power bars?

It’s going to have some background about the user and it’s going to use that to personalize it and lean into that. So truly, if you’re not paying close attention to your audience, if you don’t know who they are and how they exist in the world, you’re missing probably the most important opportunity because the LLM is hyper-personalizing its responses. So again, organic, you’re here to just trade information, provide insights and facilitate these conversations.

We want to constantly be applying this process and then editing it until we get it correct. And then once it’s correct, set it on the shelf to monitor it and move on to the next question audience pair. So that’s a lot of words I’m going to recap really quickly. So what I heard us say was each channel has a clear role centered on the same audience decision, and this leads to more effective activation by maintaining the continuity of that user across all channels.

So when they leave the LLM to go validate, they see the same story, the same facts reinforced. When they click through the citation in the LLM, they see the same story, the same facts reinforced. Assuming we pull that off, the final step’s determining if it works. So we should talk about what signals tell you that your strategy’s working and how should those signals change what you do next? So in organic, we’re going to look at presence, do I even show up? Based on the question I know my audience is asking, am I even visible?

We’re going to look at citations, not just if we’re visible, the citations might tell us a little bit about why we’re not visible. We don’t fit in this class of answers. We’re not present on these types of surfaces. Source visibility, how the brand is being represented, and whether the priority questions that we chose are actually gaining traction.

We can use web analytics, visits from AI attributed sources. If you’re using GA, it’s got the AI assistance bucket now. If you’re using that, just know that that’s under counting. We got to go in and do some tinkering with our web analytics to get all of those referrals. If you haven’t heard about dark LLM traffic, that’s a whole topic to itself, but that’s still a really good source because you can tie those visits to actions on the website and that’s going to give you a sense of how much traction you’re getting. We also use LLM visibility tools to get a sense of do the citations that we worked so hard to get included, are those sticking or is the LLM pulling something different every week? And if so, what do we do about that? Really what we’re looking for is, can I get my brand mentioned?

Can I get my sources pulled in that help validate my brand as an answer? And then am I surviving the conversation? And this is a developing area, right? The organic LLM optimization world is just starting to talk about this in the last couple of months. If my brand is mentioned at the start, as somebody goes through the process of asking questions and refining their search, does my brand survive to the end? And that’s something else organic will be able to help with as well.

Chalice, what does this look like from your perspective?

Chalice Jones (43:25):

Yeah, I mean, I think that because of the forever evolving nature of all of this, we have to constantly be looking at the directly attributable signals as well as the contextual ones. A lot of what we’ve talked about are the pieces that either validate or provide context or insight or all of these other pieces to it. So it’s rarely going to be that one-to-one. We’ve got to look at it as a whole ecosystem and what are all of the contributing factors within it. And so there is a level of creativity and innovation.

There is rarely kind of a one size fits all when it comes to is this working because it comes back to what are we trying to do to what was saying, what are we asking? What’s the role we’re giving each channel? And so for those brands that are particularly trying to. The big conversation for the last while has been Reddit, YouTube, we want to get those going because we want to show up more in the LLMs and all of that.

And while that’s all well and good, it’s not just, “Hey, did it help me show up on the LLMs? Am I doing well on that platform?” Because there’s all these, again, factors that are going to send signals because if you’re not getting the reach and the visibility and the signals even algorithmically on those platforms, then you’re also not necessarily supporting the reach you could have within the LLMs and bolstering it.

So there’s many different things that you have to look at. It’s really layered. So it’s really first some of those direct responses when it comes to social is going to be, are we seeing the higher quality metrics, as high value metrics like save shares, people talking about it, the comment quality creators, what people are saying, things we’ve mentioned. But then we’ve also got that kind of message adoption. We talked about are people reflecting back what we’ve said?

Are people reflecting back what we are trying to make our promise? Then lastly is that movement over time. It’s when we put this initiative forward, how have things changed before and after? Are we measuring that, making sure that we’re able to see that shift when it comes to considering, again, those ecosystem factors. And so really it’s just using all this to refine what it is we’re choosing and then prove it and then activate the next piece.

Josh Squires (45:59):

Inna, what about you? What do you think?

Inna Zeyger (46:00):

There’s a lot there. I’m probably going to reiterate so much what Chalice said because it’s a cover for not just what’s happening on the social side, but on the paid media side, the measurement piece has to go back to something that we just talked about in activation Chalice mentioned too. You have to give the channel a job, and that’s especially important right now with something new like ChatGPT ads.

So I know that’s been a huge topic of conversation and a lot of brands want to test it, but the basic rules of marketing don’t disappear because the ad happens to be inside ChatGPT. If its job is to reach somebody earlier while they’re still working through a problem, I’m not going to immediately compare its cost per lead to paid search or paid social where somebody might be already actively looking for a product or service or ready to make a decision.

So those channels are all doing different jobs, but I still have to hold ChatGPT or any other channel accountable to something. So the questions that Chalice posed, did we reach the audience we wanted? Did they engage? What did they do next? And ultimately, did we see qualified leads, customers, or whatever the business outcome we’re trying to influence actually happen? And ultimately this goes back to really the basics of good measurement, like incrementality.

So that’s where exposed versus unexposed audiences, geo testing holdouts, branded search lift, and sort of the subsequent conversion behavior becomes very useful and it’s a premise and concept that I think we’re all as marketers familiar with. So it’s really being able to apply that, but we’re not going to get a perfect attribution path from ChatGPT to conversion, but I don’t think that’s the bar. So the bar is what job did we give the channel?

Did it do that job? Did it add something incremental? And then really finally, is there enough evidence for us to be able to scale, to adjust or to stop? And that’s just measurement. The platform is new, the fundamentals aren’t. And just a really quick shameless plug because we are going to be having an article coming out on ChatGPT advertising. It’ll be a little bit more in depth, so keep an eye out because I know we only touched lightly on ChatGPT ads as a whole. But really the bigger point I want to make for today is that whether it’s ChatGPT or something else, I don’t want marketers to feel like they suddenly need an entirely new playbook or set of marketing principles to be able to measure their performance. And that’s my soapbox.

Chalice Jones (48:32):

Oh, good. I like that.

Josh Squires (48:33):

Well, your soapbox brings us to the end of our clear framework. Again, real straightforward. And when you get to the end, it doesn’t stop. You either move on to the next question or you go back and you refine the question you were previously working on. Any point of failure in the chain, go back to the start, work through it again, and keep processing. So before we sign off, well, we’re going to share this out with you guys. If you want to snap that QR code there, we actually have a more detailed version of this list in table form, really handy to use, especially if you’re still working through these processes yourself. This makes it nice and clear and portable. Happy to share that with you guys.

And then I also would be remiss if I didn’t talk about our Forrester report. We’ve talked a lot about audience today as a key focus for LLM visibility, not just in AI discovery. We’re underscoring the efficacy of this approach from this report by Forrester. Forrester talked with Amsive clients and found that we’re able to deliver huge wins for businesses. I think the number was like 130% over three years modeled, really fantastic stats. Feel free to scan the code and learn more about audience-led marketing in the Forrester Total Economic Impact of Amsive study.

And so we’re going to hop over to Q&A. I want to make sure we leave time for folks’ questions. All right, excuse me while I pull these up. As AI answer engines increasingly incorporate reviews and public conversations from platforms such as Reddit, Meta, Google, and other social channels into their responses, what strategies can companies implement to positively influence and manage the impact of their brand from associate generated commentary?

Chalice Jones (50:43):

Just for clarity, and maybe our behind the scenes gurus can answer this maybe in the chat, but by associate generated, the question would be, is it employee generated or a partner would be for that, but we can try to take both angles here. I’ll just jump in with that. I think that social, again, it’s becoming something that you’ve really got to consider your entire toolbox that you’ve got.

You’ve got to consider, is the audience going to resonate with hearing your SMEs that are internal? Is that the kind of product they need to hear from that validation, from all of that? Or is it that they need to hear from third parties like creators, depending if you’re in CBG, B2B, SaaS, it’s really going to depend. However, foundationally, it’s helping, honestly, at the end of the day, it’s helping cultivate those voices and making it easier for them to opt in to participating in those spaces because it can feel really intimidating a lot of times, I think, for associates to participate, but giving them clear ways and avenues to do it, which is why having foundational social, not only social.

The social playbook can’t just be for the brand any longer. It really has to consider your entire, I know I’ve used the word ecosystems sometimes, but your entire or your brand’s ecosystem between your employees, staff, partners, all those things, and considering those voices and bringing it around so that way you can leverage them in the way that you can in the best way. Hopefully that answered that question. I know we’re going quick.

Josh Squires (52:25):

Second question, what should marketers be asking our line of business executives to do or their teams to do or not do? I would say the checklist is probably the place to start. That’s a really good list of things to be doing and not doing, and we’ve outlined them for you. Be sure to grab those at the end. I would also say, again, having that messaging matrix and just knowing what questions people are asking. By line of business, what questions are relevant to which audience segments, right? I’ve definitely worked in structures where line of business owns a certain client segment, so great.

Pile all that data in, flesh out those personas really completely, go and look at the spaces that they occupy online and get a real strong understanding of what their concerns and needs are, and then go and mine those questions ideally via LLM visibility tools, although there’s some good data you can get out of Google Search Console and Bing Webmaster Tools today, answering those questions, seeing if you’re visible for them, and then using that checklist to work through the process of becoming more visible.

We have a third question, what can paid media test quickly and what needs stronger foundations first?

Inna Zeyger (53:46):

So, so many things. So I think ultimately, and maybe going back to this leads in with what the previous question was around what we should be telling execs or somebody who’s managing a line of business, you have to get clear on the business problem first around what is your actual business problem and is AI going to actually be the solution to that? But from a paid perspective, I think messaging is probably one of the key areas where you can test pretty quickly to go across multiple channels.

And I think I talked a lot about you can have five different places that information shows up and then you can then take paid to quickly test. We used to do that quite a bit with Meta to figure out what kind of messaging should go into print advertising or other channels that might have less flexibility with being able to modify the messaging, whereas with the media side, that’s pretty quick.

I would also probably look at areas where if you’re looking to have a lot more content, I would say mine your search queries also, mine the conversations that are happening around your product or brand, and then take the top volume ones and build out content on your site to support that. And that will allow you to leverage things like AI Max where you can have the landing page, you can actually have AI use different landing pages on your site.

Of course, you can control which ones it looks at and doesn’t look at to evaluate conversion rates with the fact that you have a lot more content that aligns better with who you’re targeting as well as the ads and the message that’s clicking on them as opposed to I think what historically has been, I think I see a lot of advertisers just have one landing page and maybe not enough content that supports what we’re trying to target or the message that we’re trying to get across to land them to.

So that’s really fulfilling the promise of the ad. I would say foundationally, I think you probably want to have your tracking and measurement in place, and that’s often, I think, the biggest challenge. If you don’t set up what the framework is and what your source of truth is of whether this is a lead, whether this is a conversion, a lot of that testing becomes suss or suspect because you need to have that established first.

Josh Squires (56:29):

Great answer.

Inna Zeyger (56:31):

And you can talk to me more about that. Feel free to reach out for more ideas.

Josh Squires (56:37):

All right. Well, we have one last question. For established proof, how do you strongly encourage, enable, and support clients to create truly differentiated thought leadership that is backed by original research that competitors and publishers can’t easily replicate? Yeah, that’s the hard one, isn’t it? What

Chalice Jones (56:57):

A good question.

Josh Squires (56:59):

I feel like this is really –

Chalice Jones (57:01):

Flowers.

Josh Squires (57:02):

Variable by industry.

Chalice Jones (57:03):

Yeah.

Josh Squires (57:07):

Internally, you probably have some data you can leverage first, and I think that’s where everybody goes, and that’s probably going to be the most common answer you get. And I will be the first to tell you that that’s just not true for some businesses. You maybe aren’t storing all of your data in a database that you can easily query yourself and have your data engineer come up with a dozen interesting facts about your business. Certainly easier for some than others. Personality aligned with who your audience is.

Again, the value of audience research, right? If you can’t come up with unique data points easily, the next easiest thing is to just be likable, have a personality that people want to engage with. And when we think about this, it sounds a little ridiculous, I think, but think about the brands you truly like. Think about the brands that have a lot of good will.

There’s something about them that just makes them pleasant, enjoyable, somebody you want to interact with. Chewy always comes to mind for me. I think they’re a very thoughtful company. I think their marketing is very solid. It’s not intrusive, and there’s a nice friendly tone. So I think that’s a good option.

Chalice, did you have anything to add? You look like you had a thought.

Chalice Jones (58:25):

Yeah, I think it’s exactly what we were talking about earlier too. It’s having that niche focus when you are able to own what you are understanding the audience, like you were just saying about Chewy, understanding even if you don’t have data, if it’s likability, it’s really because again, there’s a resonance, there’s a relatability there. And so when you’re able to do that, that really moves the needle so much. And so I think what I said for the first question, it’s creating channels and ways and easy, removing barriers for somebody to be able to do that from an executive POV, but figuring out what feels authentic to them.

Josh Squires (59:04):

Wonderful. Well, I would love to keep going, but we are out of time. I want to thank everybody for showing up today and spending this time with us. Thank you for the great questions. Again, be sure to grab the QR codes and check your email after the call. We’ll be sending out some materials. The recording and resources will be available. Thank you all for joining us today, and thanks to Inna and Chalice for sharing your experience with us. Thank you.

Inna Zeyger (59:30):

Thank you so much. Bye.

Explore Answer Engine Optimization (AEO): Your Complete Guide to AI Search Visibility, or let’s talk about how Amsive can help you supercharge your AI discoverability.