Joel (00:01.643)
In B2B, customer trust is not a soft metric. It's what moves deals, generates loyalty, and builds the kind of credibility that no company can manufacture on its own. Welcome to the Wings for You podcast. We are here with Joanna Borkowska. With 16 plus years in the business, Joanna is still the one your B2B marketing team can call when they're drowning in content strategy, delivery, and operations, or need a hand with websites and put their names out there in a lens.
How are you, Joanna?
Joanna Borkowska (00:33.75)
Good morning, good afternoon.
Joel (00:37.199)
So, AI is changing not only how we create content, but also how people find it. Today we'll look at how to optimize for AI without losing the thing that makes customer stories valuable in the first place: original, credible, and human experiences. So, before we get into customer stories specifically, I think it's worth setting the scene a little bit. We're hearing terms like Geo and EAO more and more.
But the distinction between them isn't always very clear. So how would you define generative engine optimization and answer end chain optimization? And where do they overlap?
Joanna Borkowska (01:18.11)
Mm, yeah, so there's a lot of mix-up with this because there is also some overlap. And to me it's kind of similar situation to what was happening when machine learning was kind of in the mainstream. Everyone was mixing up machine learning, AI, neural networks, right? So AEO is uncertain optimization, and I would say it refers more to the structuring of your content, to the more technical part, even
Even overlapping a little bit with a technical SEO, traditional as we know it, how you lay out your thoughts, how you make your content extractable, how you use the headings and FAQs, elements like that, in order for AI engines and crawlers to extract to find you and to provide you as an answer to to a question that someone's asking.
And then GEO, which is generative engine optimization, is more about PR of the AI, I would say. So how your brand is perceived and strategically where you where you put this content, what kind of associations you want to build through the content, and how you are getting your authority and how you are seen, perceived by the audience.
Joel (02:34.518)
Hmm. I I find it very interesting this PR of the AI comparison. So I have another comparison. So if we compare that with something most content teams already know, like SEO, what actually changes? How are they different from the traditional CEO approach?
Joanna Borkowska (02:52.534)
Yeah, that's a favorite question. And everyone's obviously always talking about it. But I think there's by now there is some common agreement. whether one is over another, that's still unsettled. Like even today, I saw a LinkedIn post when someone was touching upon that. Like, yes, is a CEO part of AEO? Is a CEO over like above? I would rather say let's let's treat it like that. SEO, traditional SEO obviously still obliges and it didn't lose its relevance, it just shifted. So
It used to be a discoverability tool, right? You would optimize for SEO so few people could find you. Now, when we know that 30 to 60 percent of organic search clicks decrease, there's a decrease of 30 to 60 percent, because we just type in something and we see AI overview and we tend not to go further unless we're on the decision stage of the of the buying, for example, buying stage. SEO has become a tool for A AI, AI to find you really.
So it's part of like a search everywhere optimization now, more than a separate category. So I would treat them side by side. So as SEO is still the foundation because when you think of it, most of the AI clo crawlers engines they take from search indexes, especially if they have providers like Google or Microsoft who have both search index and LLM. Then it's like tied forever. It's not gonna be separate. So if you wanna there's a
clear col correlation, like if you want to search be searched in AI in LLM tools, then you have to be also searchable and rank high in preferably in first top ten positions of search, traditional search.
Joel (04:35.436)
Hmm, that's the search, right? That's the the challenge. at Win4U we work specifically with social proof, where the customer's voice brings a kind of third party credibility that brand and content alone can't always provide. in this new landscape, what role do customer stories play?
Joanna Borkowska (04:56.427)
Hmm. So I would say, and I say it with all my responsibility, that customer stories have become like they change from the supporting pill supporting content, call kind of collateral content, you know, like marketer goes to product comp to product team and says, like, give me some case study, I need some content. It moved to pillar content, it moved to one of the most foundational elements of your AEO strategy. why is that? I would say because
if you are aware, if you're familiar with the term EAT, the principles of experience, expertise, authoritativeness, and trustworthiness, they come from traditional SEO, right? But they're it's extra, extra viable and important for AI as well, because AI engines prioritize content that's unique, that hasn't been used anywhere, that provides you like real user-human experience.
That has numbers that haven't surfaced anywhere. Anything that AI is not able to manufacture because it hasn't learned it yet. And case study is just like a definition of this, because you have something that no one else has lived through, you have something unique, you have human voices, you have authors, and you have quotes that you can attribute to a specific person. So because of that EAT principle and importance in AI search, I would say
customer stories like bang on them, but bang on them a little bit differently than you did before.
Joel (06:29.624)
So in a way it's great, right? Because trust and first hand experience and credible customer evidence matter even more now.
Joanna Borkowska (06:39.265)
Yes, definitely. And obviously, I think maybe we'll touch upon it a little bit later, but we all know like recent updates in at LinkedIn or in Google content guidelines, they clearly say do not mass produce AI content because it will be it can be also p penalized if we detect that it's just been like, you know, assembly line produced. we're looking for individual, human and unique and case studies and customer stories, they meet all of these criteria.
And also because people honestly by now they just fed up mo many people, including myself, fed up with AI content and people still read this. I know we're heading into the direction where AI and agents will be talking and reading and we won't do anything, but I don't th I don't think it's just now because if you have if you have a software company who sells their solutions for five million dollars, I'm not very
Sure, if a CEO of a buyer company will just leave it to AI, they will still want to read and check and see and fact check and see credible proof. And that's what case study is, and customer stories give you that.
Joel (07:48.013)
Hm. Yes, the human input is still certainly very important, but we'll discuss that later down the line. So if customer stories already contain a lot of the things that AI systems may value, what actually needs to change in the way we create these stories and in the way we structure them?
Joanna Borkowska (07:53.579)
Yes.
Joanna Borkowska (08:08.321)
You know, like first of all, I think the change should be strategic from what I already mentioned. So it's no longer like a retroactive exercise that you need some some content in your pipeline as a marketer and you go and check with your developers team what they're working on. This can't be that right now because it will be a waste of time. now it has to be incorporated into your marketing strategy and content strategy because it's so important and because it's so valuable. so
You have to think forward. Okay, how I can elicit reviews from customers, how I can turn them into a case studies and customer stories, even to the point that company would would have to consider check their contracts that they have with their customers if they allow for the use of logos, for the use of quotes, you know, it's v essentially important now. And what change in the structure?
And more than ever you have to rely on those quotes because they're getting extracted. More than ever you have to try to elicit some numbers and get some numbers. it they don't have to be specific numbers like two million dollar revenue or anything like that. They can be timelines, they can be before and after, just something that's always unique, that's based on real example, like real experience, and also that's quantifiable in some way.
And then obviously structure wise, there are certain rules that apply for AI content. So let's start with the fact that you wanna try to put all the content that's extractable to fill those numbers, the quotes, a little bit of comparison, maybe some depending like technology, technology box, you wanna put it first, up first, like as soon as as it's possible.
And that's because AI crawlers then and very AI rankers, they do not like read your whole content. They just go, they extract elements and also they have a cap, usually a budget cap on how much data they can consume. because it's expensive, right, for them to analyze that. So you wanna put everything you did that's most important in the fair like up front. You wanna have your snippet text, you wanna have like summary in the beginning.
Joanna Borkowska (10:20.841)
you wanna have headings that are maybe a little less abstract than they used to be. Like I love narrative headings, but now it's better to make them more concrete. Like, you know, add a number, add a name, add a question because questions rank really high. It's confirmed that you know, that you have like double chances of being cited when you have a question mark in the fragment.
But again, don't force it. It can't be like something that you just you just may turn it into a question because AI will pull it, because AI is also not stupid, it's smart. So it has to be it has to be has some reason to to exist. So for example, if your if your customer would ask you some question like, How does your software help my customer support teams? Put that in the heading and then put the answer in the first sentence that's important, and then build a pr paragraph.
And then it can be extractable because it's like it has context, it has it's relatable relatable, it has some basis for being there, right? in terms of the length, it's been researched that from 120 to 180 words, those are the paragraphs that tend to be pulled out the most, not shorter, because shorter ones do not provide enough substance.
Typically not longer, because because yeah, because AI is looking for short answers. So one twenty to one eighty, one eighty words. However, it's not again, it's not like a set rule that you have to you cannot do two hundred, you can do two hundred. If you provide the value, do ha two hundred. If you're just doing dilute your messaging, don't. Just use your good human editor to help you with that.
Joel (11:59.277)
Hmm. So a lot to think about here. but what I'm actually curious about is that we also tend to talk about AI discoverability as if all LLMs worked in the same way, but that's not necessarily true. Like we are talking here in general about AI, but how much does actually the platform itself matter? Should teams be thinking differently about how they position or structure story?
Joanna Borkowska (12:11.297)
Mm-hmm. Yeah.
Joel (12:25.986)
depending on whether they want to surface it in Chan C P T or in Clore or in another tool.
Joanna Borkowska (12:32.329)
Yes. So that's a very important consideration and you're right that obviously we put everything like umbrella turn on everything because it's just easier to to talk about it. But we go we go into details. every single LLM tends to pull data differently from different sources. And also not only that, it pulls data differently for brand mentions because it usually works on its own learned data. So its own whatever Chat GPT learned data has in its base, it will pull from that.
class from places places like Wikipedia, like industry directories, like review sites, and then for citations, they usually work with whatever they can retrieve live, like in the moment, but they also differ where they take where they take knowledge from. So one example is perplexity. Perplexity supposedly tends to rely a lot about on community platforms on reviews.
And it's an interesting thing. I can I can tell you an anecdote about that. One of my customers, a long-term customer, is in network automation, very, very specific niche kind of industry. So it's mostly technical people talking to technical people. I have never used perplexity in my life. I use other LLMs and they use it constantly. So this is this is your audience. Like they write to people like themselves who are on perplexity, who are interested in very technical content.
So they their case studies would have to be also distributed in the sources that perplexity looks into. So community platforms, so trying to get a review from the customer and tie it with a link to the case study. So that's how the marketing manager over there would have to think: okay, how do we get it in front of this particular source? obviously, copilot draws from copilot draws from Bing.
models chat chat gpt draws from Google index and open and openai. So different different platforms take from different kind of providers and different places. And yes, it's important again because of the example I gave of perplexity to understand your audience. I mean it all boils down to that in the end. Like you write a case study for the audience and you think where to distribute it for the audience. So then you should also do your research and check, okay, if I wanna my audience is predominantly on cloud, what do I do?
Joanna Borkowska (14:55.371)
Although with Claude is a little bit tricky because they they're the most mysterious and it's hard to tell where they take data from. Yes. Yeah. Yeah, I mean, maybe not anything, but like compared to other providers, this is extremely enigmatic. So this is like how and I, for example, this is my favorite LLM. So yeah, but again, know your know who you're writing for. And that also applies to the language, because I know that's been a consideration that
Joel (15:02.13)
so that's the case with Claude. They we don't know anything about them.
Joanna Borkowska (15:22.783)
Especially for B2B or software products, we often have different stakeholders, right? So we have people who like business developer and development person who just goes through different sites and different products and thinks, okay, what what we could we use to optimize processes? But then they run the same content through their, I don't know, solutions architect, who is a very technical person, right? And how to ensure all of them see see this in the in the age of AI engines.
answer to this I would say and that's also what's kind of changing in terms of how we approach case studies than than we used to before. Distribution, redistribution, chunking, recycling of your case study. So it's not like before that you wrote a case study, you just put it a LinkedIn post about it. Hey, we have a new case study and that's done.
Now, for many, many reasons, you wanna make a video snippet out of it. You wanna make five LinkedIn posts touching on you know different different areas. you wanna publish it on another site, but not copied. That's very important. Why not copied? Because if the other site has a bigger authority than yours, because of like in general, like brand authority, they will they will have the traffic, right? So you don't wanna do that. But rewrite it, repurpose it, chunk it down.
For several reasons. One, you can cover more channels. So the technical architect will get it, and also the business person will get it. Secondly, you kind of enlarge your trace on the internet and you enlarge your trace and build credibility through that and connection with the topic, whatever for whatever you want to prompt for. And third is the prompts. Different, if you have one case study, but you chunk it into different elements, different kinds of assets.
you can answer different prompts like strategically work. Okay, this one could like whoever whatever puts whatever whatever in the whoever puts whatever in the in the prompt, sorry, this one will show up and then they would read it to to the case study, right? Or my YouTube, that will answer another case study, another prompt, sorry. So this is the reason why you wanna have your content all over and you want to distribute it more.
Joanna Borkowska (17:37.525)
And also covering all languages within the content itself, because that's where I started from. you wanna write from the business perspective, but you still wanna have maybe a little technical box with technologies, maybe quota solutions architect, or or maybe have a separate paragraph that is more technically oriented, but you wanna cover more more ground, like both grounds.
Joel (18:02.432)
Ornai Ira Boss.
Joanna Borkowska (18:04.319)
Yeah. Sorry, I got y yeah, maybe. I got a little bit mixed up.
Joel (18:11.128)
it wasn't for that actually. It was because I was signed out of Google Docs. but I I didn't feel it was that long. I I feel it it was as long as it needed to be. So I just
Joanna Borkowska (18:15.787)
Ha ha ha.
Joanna Borkowska (18:25.901)
But I don't know if I can be edited out. I I got a little mixed up with the words there then there, but I don't know, up to you. Maybe you can edit it.
Joel (18:33.354)
If you want we can we can go again.
Joanna Borkowska (18:37.025)
Yeah, let's do it again. Let's do this one again. Just give me a sec.
Joel (18:39.553)
Okay.
Joanna Borkowska (18:47.317)
Thanks.
Joanna Borkowska (18:51.309)
It's just like I have so much more, but I'm like mixed up like everything, you know, all the data that I wanna touch upon. It's just so much. But yeah. Thank you.
Joel (18:53.972)
I was also
Joel (19:00.358)
you're doing great. I was thinking Orne maybe we can have since we can have her only on the screen, in the layout, we can have her only her talking when she answers some questions.
Joanna Borkowska (19:25.791)
Mm-hmm.
Joel (19:32.002)
Yeah, right. Like having only her. So I'm not nodding and s just nodding endlessly.
Joanna Borkowska (19:38.399)
Well, I don't know, Joel. I think it acts adds some because people will look at whatever whoever they wanna look, but it adds some kind of dynamic, no?
Joel (19:48.622)
Yeah, well it
Joanna Borkowska (19:49.645)
Whatever you prefer, but yeah, I think it's cool.
Joanna Borkowska (19:57.729)
Yeah.
Joel (20:04.767)
you get the video separate.
Joanna Borkowska (20:06.893)
okay. So you can like zoom in, okay.
Joanna Borkowska (20:19.979)
Me?
Joel (20:20.3)
Yeah, the thing is I don't have the my in my view you know, it's not the the camera is very up and the you guys are very down, so
Joanna Borkowska (20:38.913)
Mm-hmm. Yeah.
Joel (20:46.176)
okay. So the question was the one about LLMs, right?
Joanna Borkowska (20:53.229)
how to how to the where they take data from because I did so many detours. That's okay, let's do let's do that. Let's do that. Let me let me just check my notes so I don't confuse okay. Okay, chat the pink. Yeah, I always confuse Chat GPT with Google. So I actually made a mistake there, so yeah, let's repeat that.
Joel (21:03.041)
Okay.
Joel (21:15.167)
okay.
Joel (21:18.966)
Yeah, yeah, the sixth question, number six.
Joanna Borkowska (21:21.532)
no question.
Joel (21:25.742)
So what I'm actually curious about is that we also tend to talk about AI discoverability as if all LLMs worked in the same way. So we are even now talking we about AI just in general, but that's not necessarily true. So my question is how much does the platform itself matter? Should teams be thinking differently about how they position a story depending on whether they wanted to surface it?
in ChatGPT or Claude or Perplexity or another tool.
Joanna Borkowska (21:58.585)
yes, that's a very valid question because I think it's over often overlooked. and the answer is yes, teams should be thinking differently. maybe not obsess about it, but at least be be aware that different LLMs and different AI engines and crawlers, they tend to look into different sources. So quite naturally when you have a company that's both a provider of Surge Index and of a LLM of let's say Microsoft or Google.
They will prioritize their own search indexes because that's the database where can they draw data from. So, for example, ChatGPT relies heavily on Bing and OpenAI's own crawler. And also it tends to put a lot of attention, place a lot of attention into directories, review sites, encyclopedias, and high authority pages. My favorite example that I like to quote is Perplexity.
An LLM that I don't think I've ever used in my life because I I prefer other. However, I work with one customer who is a very niche network automation, network orchestration company, and they they basically very technical people and they sell their software to other technical people and they love perplexity. So for them, if you wanna make sure that your content is visible and you have brand authority building in the place where the audience is at.
they want to target perplexity. And perplexity tends to take a lot of searches, a lot of knowledge from community platforms and from review sites. So, for example, that means if you issue your if you publish your case study, you wanna make sure you tie it to a review, let's say on a some review platform, or you wanna somehow make it visible maybe by commenting, maybe by using excerpts of it, maybe by linking to it on community platforms, and thus show in perplexity.
So yeah, it's a very valid point. you have to check your sources, you have to know your audience, you have to know where they're at. not only in terms of where they're at on the websites, but also where they're at in the LLMs. And then know where the those LM LM LLMs take knowledge from.
Joel (24:14.028)
Mm, so it's a whole other layer, right? Knowing which LM draws from which another service or platform or search engine or whatever.
Joanna Borkowska (24:24.033)
Yes. Yes, exactly. But that's also the reason why you should change your case studies and then as I mentioned and distributed on different in different forms in different channels, because then if your audience is maybe broader and maybe if you want to target both technical people and also more business people or users, depending on what you're selling, you wanna make sure that your content appears in the places, in different places that different LLMs draw from.
So yeah, it's a it's a part of different layer. It's a kind of like additional, I would say additional dimension or task for marketers to understand that.
Joel (24:58.414)
Definitely. another thing I wanted to ask you about is that there's been a lot of discussion about whether long form content still has a place in this new environment. You talked a little bit about length of the paragraph. so when it comes to customer stories, is long form still valuable for AI discoverability?
Joanna Borkowska (25:14.381)
Yeah.
Joanna Borkowska (25:21.941)
Okay, so that's a more n a little bit more nuanced because a lot of people assume okay, since there is a cap, as I think I mentioned, a budget cap on how much content AI crawlers can can see, can can look through, it means that long form content doesn't apply. It's strong because for example, one very extensive research on Chat GPT actually showed that long for content performed better.
And that's in this per specific case. And also, you know, it's like industry specific, it depends on the format type of of your asset. However, why it performed better? Because probably because it was denser, like it provided more substance. So it was deemed as more credible and more valuable. And then the crawlers took whatever they needed from that content and they cited it. It is not true.
AI again, AI doesn't read things like top to bottom. It will typically draw from the beginnings and the end. So those are the things that you should optimize when you write whatever type of content, including case studies. Focus on the beginnings, focus on your headlines, focus on your metadata as well, is very important in your schema. And make sure you have a nice, self-contained, extractable summary. And don't
If you write if you ha if you think you need two thousand words to say something, but it's something substantial, something full of, you know, fresh evidence, a new perspective, don't be bothered by thinking, okay, no way I won't pick it up because it's just simply not true.
Joel (26:59.116)
Hmm. That's actually an excellent segue to my next question because you said extractable. and that's another term we hear we hear quite a lot in making content. So making it more extractable or chunkable. So what does that actually mean in practice?
Joanna Borkowska (27:08.64)
No.
Joanna Borkowska (27:16.799)
Okay, so I would say yeah, even I use those terms interchangeably, but they're not like extractable, you wanna do extractable, you don't wanna do chunkable. why? What's the difference? Like the subtle difference is chunkable is something, you know, chop chopped up up like a salad. and even myself in like early days of AR AI optimization, I used to do that with some content. So
I would do very brief paragraphs, a lot of headings, a lot of bullet point lists, a lot of you know tables. It's still valid because when you think of it, you want to create something that AI will be able to extract, right? Easily, like pull out easily. but also you always have to remember about the human again. Like we if you have a bullet list with 100 bullets, it just vanishes, right? It doesn't, nothing stands out. So
work out the a balance, don't chunk it up just to be suitable for AI, because then also you lose the content if the context that also AI looks for in the sense that it won't just pull out things because you made a list of it out of it. It needs some context, like some substance to it. So extractable means yes, use shorter paragraphs as I mentioned, use those headlines, use elements like a pull up
pull out forms or if you have a good quote, take it out, take it like pull it out so it can be easily extracted. However, do not do not chop it up. Do not chop it out like there's no silver bullet for that that if you will chunk up your content, it will get cited, because that's not how it works. Remember, I think a like good rule of thumb is just to remember always about your audience and always about the purpose of your content. And then you optimize it from there rather than the other way around.
Joel (29:05.368)
So extractable yes, chunkable no.
Joanna Borkowska (29:08.619)
Yes. And maybe one more thing, sorry I forgot to add. Extractable, so how to think about it. every piece should be self-contained. So able to you would be able to understand the giz, like one idea, one paragraph, one idea. But honestly, there's nothing new. This is like basic principles of good content writing, right? So there we go.
Joel (29:29.506)
Yeah, that's true. and speaking of the way we say things, language is another interesting part of this. a company might use very specific product terminology, depending on the industry and everything related to it, but while the customer describes exactly the same thing in much more everyday language, maybe. So if discoverability is the goal here, how should terms
Should teams balance those two? Sorry, I'm going to start the question again. So, speaking about how we say things, language is another interesting part of this. So, a company might use very specific product terminology depending on the industry, while the customer describes exactly the same thing in much more everyday language. We see it all the time in interviews and in case studies.
If discoverability is the goal here, how should teams balance those two? Which language should take priority here?
Joanna Borkowska (30:33.353)
Again, this will be a piece of advice I would keep repeating. Always think about your audience and always think about human, even though if you want to optimize for AI. So, first of all, AI is a smart system, right? They're intelligent, they can match the meaning. So it's the same kind of thing that we had. It's not early, early stage SEO that you had to provide exact phrases. It's late stage SEO plus more. That it can really from the context, it can read out okay, this will be interesting for them.
However, thinking about your audience, if your audience is both technical or stakeholders, let's say decision makers who ultimately lead to buying your equipment or your software or whatever services you have, if they're both technical and business oriented, you'd naturally want to address both of them. I would say in majority of cases that I've worked with, you wanna you wanna drive the business kind of logic and business language to be kind of more open and
approachable for a wider audience or marketing manager for business development person for product manager but then still instill the technical detail into it so for example invite a solutions architect to the conversation and quote them right make a little pull out box with with the technical data it kind of like a data sheet on a side so try to cater for both drive it mostly business like but try to cater cater for both in a natural way
So treat the technical detail as an addition and then of course depending on your audience, b again, because if you're like strictly technical and it's only technical, just do that. But do not try to, you know, cater for everyone, try to cater for whoever you speak to and you'll be you'll be good because AI will do the matching part as long as you provide enough of valuable information in the background.
Joel (32:26.038)
Yeah, it it seems more and more that personalization is or customization is really the thing to look after for here. and I I have another question also regarding the structure of the customer stories or the features of the customer stories. So metrics have always been a big part of them. but how important are they in this new AI driven context? I I I'm imagining
it's very important, since we talk about, you know, making them very unique and very personalized. But I wanted to hear from you, what do you think about it or or what evidence do you have around it?
Joanna Borkowska (32:56.765)
Yeah.
Joanna Borkowska (33:07.789)
So from the research that I've been looking into, if you have a lot of credible, unique and kind of first time quotations, statistics, metrics, data points, it can raise your stakes of being quoted by AI engines like from forty to fifty percent. So basically double double the yeah.
So, yeah, it's still important, even more so, because that's what AI fishes for. But again, don't obsess about just you know stuffing your stuff, your your content with data. Do it naturally. Get those data points, get a few of them. quality over quantity again. But yeah, this is one of the most important evidence levers in terms of optimizing for AI. So I think it also is important to change maybe the way again,
from contract, like starting from contract with your buyers, to make sure that you're able to share those get and share those information to the to interviewing the customer, to really try to drive them into into some metrics, comparisons, quotes, nice quotes, you know, elicit some nice quotes, because it does make a difference. Because that's the on again, that's the only thing that AI cannot really fabricate because it has no former knowledge of it.
Joel (34:29.31)
Mm. And getting metrics out of a client, as you said, is or a or a customer story is one of the hardest parts of an interview. so as an interviewer I I should know. But I I was thinking, does all of this change the way we should be conducting customer interviews too? Is there something that changes there? Or just I'm curious because it's my job.
Joanna Borkowska (34:55.597)
Yeah, and I I'm just thinking if that's good news, that's good news for you or bad news for you. not really, in the sense that if you've been doing your interviews right from the start, nothing changes really. Because if you've always been trying to get those quotes, if you've always been giving a lot of grounds to your the spokey speaking person to share their story, maybe yeah, maybe in this way, just
nudge them a little bit more to share whatever is personal because even things that you might not think that will be valuable in the end, like very personal things like how their life changed because some system was implemented. Very, very like individual stories. I think they are gold now, because that's again something that cannot be that's that's theirs, right? And that's relatable and that's something refreshing. So maybe just it didn't change the way you have to interview, but maybe put emphasis, more emphasis on certain things.
And also I think it's worth thinking about the backstory more and maybe highlighting, emphasizing the backstory, not only the problem, but also how we found the provider of the problem. Because what I know, and again, like having done like hundreds of case studies, I I also I'm also guilty of that. We also always do like before and after, and then in one sentence we put, and this company X found company Z, and then then we go, the solution is was there.
So it's like one sentence story about how they found it. Whereas the procurement part, you know, like kind of struggling where to find the supplier or how to choose them, it's actually quite interesting too. And maybe it doesn't add that much to it's not that it increases your chances of getting found by AI, but on the whole, it increases the individuality, scarcity of the story. So it's kind of something also worth considering.
Also during the interview as much as the for the content writer who makes the story, to expand on that a little bit.
Joel (36:51.534)
Thank you. I'll make sure to take a note out of it. So there's also a bigger question here around the role of humans, of course. So with so much focus on AI optimization, where should human writers, editors and interviewers and customer teams still be taking the lead and how much of the process actually makes sense to hand over to AI tools?
Joanna Borkowska (36:55.233)
Mm-hmm.
Joanna Borkowska (37:19.989)
Well, I won't give you the numbers. We all know that last year, I mean people in content and marketing know that last year was like the year of the comet, like everyone was just poo, getting rid of content content writers and editors and then replacing them with AI. Obviously it backfired because of some of the things I already mentioned. So already some companies started penalizing either mass AI mass produced AI content, like clearly.
Or content with qualities that are usually attributable are attributed to very lazy AI writing. So you you you just prompt the AI to okay, speed it out, like I I give you transcript and speed out that case study. So this is going to be either lowered in ranking or even ignored by AI crawlers for several reasons. First of all, again
By nature of it, AI eng engines are not creative. They are kind of they regurgitate whatever they already know. second of all, even if you feed them with original data and with those transcripts, let's say, they lack creativity also in the in the layouting, in the writing process. We all know I'm not even talking about, you know, the overuse of actually unjust and whether this then that.
Joel (38:18.158)
Mm.
Joanna Borkowska (38:36.459)
I'm not but those are also kind of patterns that AI algorithms that fight with AI writing they they discover. However, I'm I'm more talking about the cadence, the flow, the rhythm of the language, the kind of ideas, how to structure the story. It's very limited. When you when you think of it, when you read enough, you you will just easily spot it. And this all is treated as something secondary, something derivative.
And the prioritize by AI engines. So then if you have 10 boiler to plate stories, even they if they read well, right? They read well, but AI engines will pick up, okay, this was probably machinized, like this was generated like that by machine. I'm not sure of its value, I'm not sure of its human component. It's gonna deprioritize it and prioritize whoever has the biggest authority, like globally as a brand. So then you cannot really compete with other brands for a prompt because
They already have bigger author authority and then you provide content that it's kind of the same. So you cannot differentiate yourself. So the human p component, human judgment, human creativity and skill are what you can how you can stand out, right? so that's another thing. And I'm not even talking about, you know, hallucinations. And obviously we had very recently also very specific Google guideline about it. Fact checker
facture your stuff basically like don't let AI just fabricate things and honestly I'm just mind-blown how how bad it is working with AI extensively. I'm like it's error-ridden in terms of you know manipulating and manufacturing facts, quotes. So be very careful because if you plan to just take a transcript, put it into AI, prompt it to do something, to write something for you.
It will even make up quotes. It will just it was just add things that don't exist. And this is penalizing, this is confirmed. and it's I think the trend will be that it will be penalized more and more. And again, from the point of view of your audience, just imagine if you're like a I don't know, business development person who has to evaluate 20 solutions. Like you don't want to read, you don't want to go through tons of tons of boilerplate, Dow.
Joanna Borkowska (40:57.317)
humdrum things, right? It something will stand out and something will touch with your touch with your like be relatable and something has a human tissue that will stand out and then you it you it will make you memorize it more, right? And then it will make you think, okay, let's check out this company maybe. So I don't know the percentage value of how you should how much you should AI, how much you you you shouldn't. I never use it for drafting. I never use it for outlines because they're just very repeatable and they're very
bad quality honestly. I use it to, for example, adapt the tone. I use it to brainstorm as sometimes like the storyline when I'm thinking, okay, how I could what what kind of thread I could add here or you know, like how I could spin the story. Yeah, I'm talking to it like that, like creatively, kind of evaluating my own ideas. But and then spelling obviously, yeah, spell check it with AI or look for synonyms. You have tizers for that, but you can do that with AI.
But obviously now increasingly, fortunately, AI produced content is downgraded and it doesn't give you a an advantage, you can be penalized if anything.
Joel (42:09.644)
Hmm. I love hearing your experience using and working with AI. this also makes me think that conversely to what one may think or might have thought in the beginning, the role of the editor or the role of the content team is way more important now in this landscape. But also there's a lot or another whole layer of work for these teams. So there's the writing itself.
but there's also structuring, optimizing, testing, monitoring and updating the content over time. So how much additional effort does GEO realistically require? And where can working with a specialised partner like Wins4U can help make that manageable?
Joanna Borkowska (43:00.449)
Think we've fallen a little bit in a trap. that's my personal experience and also experience of other marketers, where we think, yes, AI makes things faster a little bit, some things, because honestly, on some other, I I didn't find it faster. I just treat it as a tool. Like again, I have pizarros, I can put that in AI, and it doesn't save me any time, honestly. But it does save time on some things. but that means that people have more and more tasks to do. So you have to be a marketer and you have to be
thinking about strategy. Sometimes you're expected to write the content, they publish it, to analyze it, you know, optimize it. It's just too much work for one person. So I think it increasingly good marketers will do their strategic thing, the market development, market analysis thing, they will work with content marketers on strategy. They will probably need a hand with maybe content calendar, how to
how to strategically create a content that will show up for certain prompts and that will build your authority on on the online to for brand mentions. They they can do it alone. So they will increasingly have to either rely on like fractional support or hand over some part of it to external agencies because it's just too much. I cannot tell you, you know, how many hours or how many like quantitatively I won't give you that. However, just for sure it's it's more.
Because it provides you with more opportunities, but to grab those opportunities and to be visible everywhere, you have to have some support. Because you can you cannot do it all your own.
Joel (44:42.444)
That's one of the biggest takeaways of this interview, I think. So thank you for that. to wrap up, how do we know if any of this is actually working? I mean, what can customer advocacy teams realistically measure today when it comes to AI visibility?
Joanna Borkowska (45:00.297)
Yeah, so this is a little bit of a tricky thing because it's unlike in SEO where you have some kind of steady objective metrics. AI is very AI visibility is very difficult to measure because yeah, there are tools that I'm also using for my work for doing AEO audit, for showing companies, okay, you you rank for those pr you show up for those prompts and your competitors rank for those prompts or show up for those prompts. but all of this is more of an assessment.
And it's in the in-time assessment rather than steady fixed KPI. Because for example, even you can run the same prompt in the same LLM in the morning and in the evening, and once your company will show up, and then the second maybe not. why? Because of a different different factors. LM version and LLM type, your personal kind of data that AI can access,
From your even from your computer, depending on the settings, on just randomness, on the fact that AI is a predictive engine, so it predicts, it doesn't, it's not like knowledge base in a way, per se. So this is a little bit randomized. However, obviously, if you have a strong brand all over, and if you do your work strategically and think, okay, I want to try to
rank to show up for five prompts this this month, let's say, and work on it and write the content, distribute it, chunk it up. it will you will get some numbers from those tools. They will be, however, I wouldn't, you know, die for them. It's not like when it shows you 43% of visibility in chat GPT, that is forty three percent. Let's let's do a range of ten percent more or less with the a c error coverage, right? So I wouldn't die for it, but there are some elements that you can check.
such as your brand authority online, such as CTR from coming from AI engines, that you can look into. But it's still, I would say that's work in progress. I'm not sure if it will ever be quantifiable, to be honest, because I'm not sure if the providers will be willing to share this data so granular, like for example, Google does, right?
Joel (47:17.696)
Okay, that's been pretty interesting. I I feel like I I've learned a lot. My brain's got a big a lot bigger since the beginning of this conversation. So thank you, Joanna, for joining us and during your insights. where can we find you for the people watching and listening?
Joanna Borkowska (47:24.077)
Ha ha ha.
Joanna Borkowska (47:36.333)
well, we can find me in Costa Rica. Or you can find me on LinkedIn. I guess that would be the the smartest thing to do, just connect with me on LinkedIn. however, I wanted to add something. I wanted to share something as well if I can. Because some time ago I was doing like I was checking a marketing roundup of some experts in SEO mostly, actually, traditional SEO, and I found a quote and I wrote it down because I really, really like it, and I think it's really relatable and a good thing to remember.
So someone one of the experts there said asked about the strategy for AIH in content. He said, start with the audience, tell a strong story, then make sure it can be found wherever your audience chooses to look. And it's so universal and it's it seems like a blanket statement, but it's not because like in marketing, you always want to start with the audience. You wanna have a story, especially now it it has to be strong. So again, unique.
something that never surfaced before, something that AI cannot manufacture because it doesn't have this knowledge, right? Something attached to a person, attributed to a person. And I forgot to mention that that it's very important to have your names out and authorit authors out and also bios out. So that's also a reinforcement signal. And then make it make sure it appears wherever your audience looks for it. So yeah, check where LLMs take their knowledge from, take their sources from.
make sure you're still on the just human human side kind of like make sure you're still on the media, on the channels that your audience is at. And always think about that. Always have this in mind rather than obsessing about whether I have five questions or FAQs in my stories. So yeah, that's I just want to share that because I really think it's a it's a good one.
Joel (49:23.416)
Yeah.
Joel (49:28.344)
Thank you for sharing that. It's it's pretty insightful and I I take away this, right? Start with the audience but also kind of end up with the audience in a way. It's just think about the audience would be the the biggest takeaway of of our conversation, I think, as well. so as AI continues to change how people discover and engage with content, I think that one thing remains clear is that authentic customer experience are still at the heart of building trust.
So the challenge now is making sure those stories reach the right audiences, whether through traditional search or AI powered tools. So at Wings4U we believe that understanding how trust works is just the beginning. What matters is knowing how to put it to work. and thank you for listening and we'll see you in the next episode of the Wings For You podcast.
Joanna Borkowska (50:20.577)
Thank very much, Joel. Thanks, everyone.
Joel (50:23.384)
Thanks, Joanna. See you.
Joanna Borkowska (50:24.429)
Thanks.
Ha ha
Joanna Borkowska (51:04.749)
I don't know what to say all ever, honestly.
Joel (51:05.238)
Right.
Joanna Borkowska (51:41.421)
No, no, it's not that. I just think I just think if it's if it's like, you know, genuine and credible because if Joel introduces me, then what do I say? Well I you know, like I say hello and then I introduce myself maybe. Let me think, let me think.
Joanna Borkowska (52:15.339)
yeah, maybe yeah. Okay.
Joel (52:17.838)
Okay.
Joanna Borkowska (52:25.299)
Yeah, yeah, yeah. Yeah. I wasn't expecting it to be thirty hour thirty minutes to be honest. I I thought it would be longer anyway, so no worries.
Yeah.
Joel (52:41.123)
Yeah.
Joanna Borkowska (52:44.193)
No.
Joel (52:50.413)
Yeah, right.
Joanna Borkowska (52:54.667)
But so what what should I what should I say with the wrapper?
Joel (53:09.395)
but so we will we will be at after she says the quote.
Joanna Borkowska (53:15.372)
Yeah.
Okay.
Joel (53:18.52)
So I thank her.
Joel (53:38.392)
Okay.
Okay, okay, okay.
Okay.
Joel (53:51.257)
Thank you, Joanna. That was really insightful. I think it's it's one of the main takeaways of this whole interview. Start with the audience, but also end up with the audience in a way. so thank you for sharing that and and thank you for all your knowledge and you know, sharing it with us today.
Joanna Borkowska (54:11.561)
Yes, thank you so much for your time as well, for all your questions. Good to see you always. And thanks everyone for listening. And again, if anything, you can DM me on Link LinkedIn.
Joel (54:24.68)
ooh, should I say that again?
Joanna Borkowska (54:26.571)
Was that okay?
Ha ha ha
Joel (54:40.288)
Okay, okay. I would I
Joanna Borkowska (54:43.821)
Okay.
Joel (54:45.344)
I was wondering if we did the LinkedIn parts or not and I I got confused.
Joanna Borkowska (54:48.46)
Mm-hmm.
Joanna Borkowska (54:52.051)
Sorry, Joel.
Joel (54:57.528)
Okay, okay. Okay. Thank you, Joanna. At Wings For You we believe that understanding how trust works is just the beginning. What matters is knowing how to put it to work. Thanks for listening and we'll see you in the next episode of the Wings For You podcast.
Joanna Borkowska (55:16.525)
Okay, now the difficult part. Do the intro. Hi Joel, how are today?
Joel (55:20.95)
Joel (55:25.134)
So I do the intro again and I ask her how is she but also what has she been doing lately?
Joanna Borkowska (55:34.273)
Wait, but I have to think what I've been doing.
Joanna Borkowska (55:39.605)
Like professionally or What I've been doing my god, I've been doing so much, but but
Joanna Borkowska (55:54.829)
Mm like now mm
no, no, no, no, no. It's good. yeah, maybe Joel, if you can just ask me like how have you been lately? And then I can say something like busy as always, October is always like quite intense in in the market, yeah.
Joel (56:14.222)
Okay.
Joel (56:18.382)
So how have you been lately? Okay.
Joel (56:31.916)
Okay.
Joanna Borkowska (56:38.849)
Ha ha ha.
Joel (56:39.822)
Yeah, yeah. I focused way too much on what I have to say and I lose my warmth.
Joanna Borkowska (56:46.293)
Mm-hmm.
Joanna Borkowska (56:59.149)
Mm.
Joel (57:00.502)
Okay, so I start at the beginning again.
Joel (57:05.976)
From the top.
Joel (57:10.326)
Okay, the hard part. Now I'm going to laugh. Wait a minute.
Joanna Borkowska (57:17.546)
Yeah.
Joel (57:23.227)
sí, cierto.
Joel (57:27.853)
Right.
Joel (57:32.664)
Sorry, I'm writing it because writing it down because I need to
Joanna Borkowska (57:34.093)
No.
Joel (57:42.434)
Okay.
Joel (57:49.409)
In B2B, customer trust is not a soft metric. It's what moves deals, generates loyalty, and builds that kind of credibility. No, I fucked up. Okay, I'm going again.
In B to B, customer trust is not a soft metric. It was moved oof. I think I was I'm more nervous now than I was before, but I
Joanna Borkowska (58:15.116)
Ha ha.
Joel (58:31.895)
Okay.
Joel (58:35.672)
Hey everyone, I'm Joel. I'm a writer and editor at Wins4U. Welcome to the Wins4U podcast. In B2B, customer trust is not a soft metric. It's what moves deals, generates loyalty, and builds the kind of credibility that no company can manufacture on its own. We're here today with Joanna Vorkowska. With 16 plus years in the business, Joanna is still the one your B2B marketing team can call when they're drowning in content strategy, delivery, and operations.
Or need a hand with websites and put their name out there in the Lens. Hey Joanna, it's so good to see you. How have you been lately?
Joanna Borkowska (59:13.473)
Joelle, first of all I have to commend you for your Polish pronunciation of my surname. That's very good. yeah, I've been busy, busy because October is usually, you know, in marketing it's always a hot period. So good, also including optimizing some websites of my clients for AAO. So yeah, and I'm very happy to be here as well. Welcome everyone.
Joel (59:37.016)
Thank you, we're very happy to have you as well. So AI is changing not only how we create content, but also how people find it. Today we'll look at how to optimize for AI without losing that thing that makes customer stories valuable in the first place. Original, credible, and human experiences.
Joanna Borkowska (59:57.313)
And that's a cat.
Joel (59:57.686)
That's it, I think.
Joanna Borkowska (01:00:02.109)
Jeez.
Joel (01:00:04.055)
Yeah.
See you.
Joel (01:00:13.838)
See, yeah.
Joanna Borkowska (01:00:30.389)
No.
