Why the website is dying | James Ward, Head of Digital Experience @ Personio

Show notes

The website used to be the shop window. Now buyers do half their research inside an LLM before they ever land on it, and the role of the site itself is starting to shift.

My guest today is James Ward, Head of Digital Experience at Personio, one of Europe's fastest-growing HR tech companies, where he leads web strategy, SEO, content and CRO across multiple international markets. Before Personio he spent years at Salesforce and ADP, so he's seen this shift play out from inside both enterprise giants and a fast-moving scale-up.

We dug into what happens when your buyers already know everything about you before they hit your site, why your own Wikipedia page might be quietly undermining your positioning, and whether websites are heading toward becoming databases for LLM front ends.

What we covered in this episode

  • Why Personio's Wikipedia page was quietly miscommunicating their ICP to LLMs, and how a single line of off-site content can undo your entire positioning strategy
  • The reverse-prompting exercise every marketing leader should block out half a day for before buying any AEO tooling
  • Why "dark AI traffic" is forcing a rethink of the website itself, from curated navigation paths toward prompt-based, deeply personalized experiences
  • The shift from technical SEO hacks to brand narrative and reputation management as the real AEO lever
  • Why Reddit hype has already peaked, and where first-party content and third-party listicles now sit in the AI visibility stack
  • How AI is turning specialists back into T-shaped marketers, and why James thinks everyone (yes, even content writers) should be building with Claude Code or Lovable
  • Why traffic is becoming a vanity metric at Personio, and what they measure instead in an era where LLM citations, branded organic and direct all blur together
  • Whether the future of B2B SaaS buying happens natively inside ChatGPT, and what that means for the role of your website

▶ Let's connect! 🔗 Niklas on LinkedIn: https://www.linkedin.com/in/niklas-buschner/ Radyant on LinkedIn: https://www.linkedin.com/company/radyant/ James on LinkedIn: https://www.linkedin.com/in/jameswarddigital/ Personio on LinkedIn: https://www.linkedin.com/company/personio/

Show transcript

00:00:00: For me, my passion is always laid within the world of digital experience and building new experiences trying to kind of always be at the forefront if what those experiences look like in how they're evolving.

00:00:10: The website is the shop window versus all our competitors although... Increasingly, we see that shop window starting to diversify out into other channels.

00:00:19: So how do you see that basically changing?

00:00:21: How we have to think about the website and work with a website?

00:00:24: this is doing my phone kind of most interesting at the minute.

00:00:27: actually keeps me awake at night.

00:00:29: The most is thinking what role in the future?

00:00:32: What does it become?

00:00:34: before we dive in your listening to the masters of search podcast?

00:00:40: Each week, I sit down with some of these modest people around the world in SEO and AI search to bring you their strategies mental models.

00:00:48: And top pieces of actionable advice.

00:00:50: if You enjoy this podcast don't forget to like and subscribe.

00:00:54: then follow it In your favorite podcasting app or YouTube?

00:00:57: It helps us get top-notch guests and create The best possible content for you.

00:01:01: let's dive into today's episode.

00:01:03: most conversations about AI search right now focus on visibility and rankings.

00:01:08: My guest today is thinking about something bigger.

00:01:11: What happens when your buyers already know everything about you before they ever visit your site and what that means for the entire digital experience you built?

00:01:21: James Ward, head of Digital Experience at Personio one of Europe's fastest growing HR tech companies where he leads web strategy SEO content and zero across multiple international markets.

00:01:32: In this episode, we'll dig into what he calls quote-unquote dark AI traffic.

00:01:37: While your own Wikipedia page might be quietly undermining you positioning why BtoB buyer expectations are about to fundamentally shift and whether websites as we know them heading toward becoming databases for LLM front ends.

00:01:51: Welcome to the podcast, James.

00:01:53: Hey there Nicholas!

00:01:54: Great to be here and super excited to talk with you today.

00:01:58: Yeah thanks so much.

00:01:58: I think There's SO MUCH in it And already from our pre-talk...I felt like there are so many ideas that i want get your perspective on but we have to give our listeners a little bit of your personal background.

00:02:13: So how did you end up leading digital experience at Personio?

00:02:19: The million dollar question and yeah, really ended up here as a result of working in combination demand generation.

00:02:30: And broader digital marketing and web products ownership for kind of twelve or more years?

00:02:40: For me my passion is always Laid with in the world of digital experience and building new experiences trying to kind of always be at the forefront of what those experiences look like, how they're evolving.

00:02:52: And you know for me joining personia was a pivot away from some other much larger organizations I'd worked with historically.

00:03:00: so i came from salesforce prior to personio.

00:03:09: Billion dollar organizations like adp who pay one in one in ten americans.

00:03:14: so it's really a shift into the business where.

00:03:18: There is much more speed there's much more autonomy and we can do much more experimentation.

00:03:22: i move much faster which obviously given that, we all need to be thinking.

00:03:33: And would you agree that the digital experience title is fairly new?

00:03:37: because I checked on LinkedIn how much head of Digital Experience there are, and it couldn't find like that many?

00:03:44: So can you give us a little bit about what digital experience?

00:03:49: so obviously everybody could do something with words but... What's concept behind this?

00:03:58: i would say fairly new within the btb space but actually it's a term that's been around a little bit more in, And the reason for that change within my role was really because as we think about how the experience is evolving, and how the website is evolving.

00:04:38: The role of the website plays... How buyers are increasingly more aware in attaining to AI and attending other channels.

00:04:44: just start doing a lot of research before they even hit your web site.

00:04:50: It has bearing on what it looks like.

00:04:54: You know, I didn't want us to be stuck in the world of just thinking about a website because actually for me.

00:05:01: For my team we need to be thinking much more broadly than that so it's more to shape us for future reports coming next.

00:05:07: then let's tap into this topic already The web site topic.

00:05:11: So how important is the websites still for persona today and what do you think?

00:05:15: How important will the Web site be in the future?

00:05:19: Incredibly important, you know.

00:05:20: for us It's its still our biggest growth lever.

00:05:23: And the website The shop window versus all of our competitors although increasingly we see that shop window starting to diversify out into other channels.

00:05:33: you know there's a lot about the channels now that.

00:05:36: Bias are able to or prospective buyers were able to go too before they, Before they reach your website.

00:05:41: so and still very important still very much the source if truth for how we want it?

00:05:48: to represent ourselves and talk about ourselves in the market as well.

00:05:52: So, in terms of future importance I don't see that changing.

00:05:56: if anything i see the website becoming increasingly important even If The Role Of That Website Changes.

00:06:01: And what do you think?

00:06:02: how will it change with basically LLMs or AI chatbots being maybe more a go-to interface people interact With?

00:06:12: um, the pulse content from the website.

00:06:16: So how do you see that basically changing?

00:06:19: How we have to think about the Web site or We Have To Work With The Website?

00:06:22: Yeah this is- This Is The Thing I Find Kind Of You Know Most Interesting At The Minute And That Actually Keeps Me- It Keeps me Awake At Night.

00:06:29: The Most Is Thinking About What Is That Role In The Future?

00:06:32: and You Know What Does The Websites Become?

00:06:34: If I Think Even About The Shift And Evolution That We're Seeing in A Lot of Sites Now Where... It's becoming the norm to see a prompt box on home pages where people are straightaway jumping into building.

00:06:49: I think we're naturally starting to see rises in things like AI chatbots, we've seen rises in LLMs and their ability to do agentic browsing at our behalf as well.

00:07:00: so naturally then thinking about what that means for the website experience and user expectations when they come to the web site?

00:07:05: For me it means that you gotta start thinking how those user expectations are revolving and what we then have to build for that.

00:07:13: So, you know do we still push users through a curated path?

00:07:18: Do things like a website navigation need to exist in the future or actually do we lean more into kind of prompt-based searching and more heavily intersearch?

00:07:27: is there... A world where we should be getting much more personalized and providing a much deeper level of personalization?

00:07:36: You know that the kind of typical playbook is btb sass which we're gonna build some pages, gated assets.

00:07:44: We are going to point you a product page and ultimately hope your fill in request form for demo.

00:07:50: then naturally ways should be thinking about how we curate our experience much more heavily?

00:07:56: And now even more heavily in the BTC world.

00:08:00: Spotify have just launched their ability prompt-based personalization as well.

00:08:06: And that is going to become the norm, it's what people are gonna become used to and we see search crews on Google expanding and the type of queries people are searching for a much deeper.

00:08:18: there's more depth generally to content discovery so We have to be prepared.

00:08:27: And

00:08:31: do you think there are things we have to unlearn when we think about the website experience?

00:08:35: So if I'm thinking about SEO, what we had to unlearning was basically over-focusing on keywords and like key words being the predominant factor for ranking.

00:08:44: Then we moved to topical coverage

00:08:46: etc.,

00:08:46: now.

00:08:47: obviously they're other changes with AI search but Do You Think That We Have To Unlearn Principles Like How We Designed And How We Basically Conceptualized The Website?

00:08:56: so i Was Thinking Maybe It's A Stupid Idea But Do we maybe have to write content in Markdown?

00:09:02: In the future, if an agent is like from now.

00:09:09: We look at the traffic split by mobile desktop tablet and then in the future will get the traffic spread by Agent & Human And The Agent can understand Markdown better.

00:09:18: So Can you give me a little bit of your thinking about stuff that might hold true The longest and that we might have to unlearn for the future.

00:09:27: If you're thinking about it, the immediate term SEO in conventional SEO I don't think is going anywhere right like still proportionally if you look across And the split of LLM traffic versus Conventional SEO traffic on your Google's of the world, okay?

00:09:43: It's still heavily biased in conventional seo obviously but with naturally yeah You know i think comes a shift the type of content and as i mentioned you know that the depth of content we have to we have two curate and think about but also.

00:10:01: Not just a content on site with the content exists offsite end quite often, in the traditional seo world that's been limited.

00:10:10: so things like back linking and being able make sure that weekend and secure links back from hide domain authority websites.

00:10:17: where is now?

00:10:18: we should be thinking much more abou.

00:10:23: how are we showing up in search?

00:10:24: generally, not necessarily.

00:10:25: How will be showing up?

00:10:26: In search for our website and for a specific domain so that encompasses things like listicle websites and third-party review sites even Wikipedia pages as you alluded to earlier right.

00:10:39: I think there's a real need to think more holistically about SEO and content optimization across the board.

00:10:48: i would then also say two point around keywords.

00:10:51: Definitely, you know starting to shift away from just pure play conventional keyword targeting tool.

00:10:57: So look at how we can focus more specifically on that kind of interconnected network is topic clustered content around particular focus areas and that again could encompass one specific page.

00:11:13: it can encompass multiple pages but if probably also covers areas the maybe wouldn't have considered previously.

00:11:19: so things like The the about us section of your website or you press release pages all some of that kind legacy content exist on your site might actually be negatively impacting your studio and your other than visibility now.

00:11:31: And so things like content pruning increasingly important being able to, To make sure that the narrative we spin across All those different pages tells same story.

00:11:43: if talking the smaller end of the market and your website proposition doesn't all reflect that, then you're going to confuse SEO.

00:11:52: You are gonna confuse real LLM.

00:11:53: so you just have think more holistically about how those things stitch together.

00:11:57: Speaking off brand narrative or like narratives in general on messaging I Think it's a fundamental shift where if someone was for example using Google Or whatever search engine they were searching For something And They Were coming To our site We basically had full control over what they see In The messaging we use.

00:12:14: Now you Have this layer Before that, which could be AI overview on Google.

00:12:19: Which would be AI mode?

00:12:19: Which is basically a curated response based on the own content we have in our site and based on third-party sites.

00:12:29: So how do you think about that?

00:12:30: Have you significantly shifted resources being time people tooling whatever towards Third party sites And what can share your approach specifically?

00:12:46: you know, just coming back to that point I made earlier around that holistic view be it on-site or off-site.

00:12:51: And then firstly comes with starting with trying to understand what that perception is and sometimes yes there are tools out there that can help this the kind of peaks, the profounds et cetera over the world.

00:13:06: but... The most logical starting point actually a lot where we started was through Reverse prompting and through trying to kind of go through those lm's in the google search platforms.

00:13:18: so understand if we searched for x. So ultimately you know how do i want to be perceived?

00:13:23: what is the result that comes up with a people coming out from one of the sighted sources that are coming up, and not where u uncover then some other areas there perhaps you would have overlooked previously.

00:13:32: so mentioned your legacy.

00:13:34: press releases or legacy wikipedia pages etc.

00:13:36: refer to you in particular way.

00:13:40: Sense of okay, so how are we being perceived versus?

00:13:43: How do you want to be perceived?

00:13:44: and then from there it's about as You say just aligning that focus on the resource to try and bridge some of those gaps.

00:13:52: So We absolutely looking into more Of the offsite work in representation than previously.

00:14:03: things like Wikipedia pages probably would have been left And perhaps a brand team or an internal comms team kind of in days gone by, but now is part of our overall content strategy.

00:14:14: If we think about the offsite review websites and although they've always been important to quality of those reviews that types have reviews that were capturing all paint a picture for us so that's increasingly important also.

00:14:30: We even looked at doing experimentation with things like Reddit communities in some of the early days as well, when everybody jumped on that bandwagon and suddenly Reddit was full of spam from AI.

00:14:43: But that in-and-of itself is another learning.

00:14:47: So I think always just trying to learn, always try into test new mechanisms but... The other thing The results change daily, literally.

00:14:58: We've monitored the same sort of prompts and responses that come up in Google AI overviews for example.

00:15:05: it can change on a daily basis so you're always having to learn and test and optimise.

00:15:10: I think thats another shift from days of SEO optimisation.

00:15:18: now is that You changed something.

00:15:20: an LLM could pick it up instantly whereas previously used submit a page for index into Google and then sit in monitor SEM rush obsessively, search console to see whether it had been picked up.

00:15:32: Whether you'd have the traction that you wanted whereas now you can obviously some of that overnight.

00:15:37: so learnings are much higher proportionally which is great for us.

00:15:41: Can we share specifics about Wikipedia discovery because this I already found intriguing when talk before hit record button?

00:15:49: Yeah, yeah of course.

00:15:50: So I think the thing that was really interesting for us there is depending on some other markets.

00:15:56: you search in and... Some of the LLM perception of personia where we were geared up to serve small renders market so very much SMBs but they're out here.

00:16:12: is this though?

00:16:13: We actually served a lot broader businesses than our sweet spot.

00:16:18: core ICP is actually within the mid-market, typically.

00:16:21: And a lot of reason for that perception in LLM platforms was because one line with our Wikipedia page.

00:16:29: so there's few things we took as learnings.

00:16:33: just when you think about how kinda correlates across some of the other types of pages that you may have out there or people may have created talking about your comparison websites.

00:16:51: The dreaded competitor comparisons all listicles as well, All of that content is being used to shape how LLM's perceived.

00:16:59: so it could only take one competitive article on a third party listicle to say X brand is designed for Ex part of the market to undo a lot of your broader strategy, particularly if it's a publication that is held in high regard.

00:17:17: So its really important again.

00:17:19: you know kind of look holistically at how are being cited and referenced by a lot those other third parties as well.

00:17:25: Now Wikipedia obviously super sensitive topic because has gained very good reputation but also led to it being fed into the training data of LLMs due to moderation and community-led approach that not everybody can just go through a platform, say what they want to have said about themselves.

00:17:46: So how did you approach?

00:17:48: where are looking for other sources?

00:17:51: That would balance line more or share something.

00:17:56: if people listening also operate in bigger businesses like Personio And they have Wikipedia pages and might also have a legacy or some sort of untrue perception that is shaped by that.

00:18:12: Yeah, I think you just touched on it there.

00:18:15: Wikipedia obviously has built their trust in for the most part authenticity and impartiality.

00:18:25: so you have to maintain kinda sales listing page with a load of marketing content, it has to be factual and actually based on reference back to cited sources.

00:18:40: So that's typically what you can do when you point at things like even some your own press releases for example as cited source is talk about credibility and maximise the credibility of points they are making its still all subject For review, you know.

00:19:00: typically is actually better to get a third party or someone impartial.

00:19:04: To submit those responses for you because it's all heavily moderated by third parties.

00:19:09: so I think that the good thing cause ultimately stops as you say everybody just going out and being able to build whatever type of content they want.

00:19:20: us sit on Wikipedia.

00:19:21: but Yeah, generally speaking being able to reference some of those existing cited sources you've got if you done internal research or a few published white papers and reports that.

00:19:30: Talked only at the point still claims it trying to make so pressure leases all about sort information is publicly, publicly visible and are all things they can use to ultimately corroborate the points he tried making them.

00:19:44: Now let's look at the perspective of the buyers a little bit more because we thought about, um... The AI overviews and-and the chativity responses basically making it I would say maybe more challenging or at least more.

00:19:58: Um like you have to spend A Little Bit More Time on thinking About How To Shape Your Brand Narrative Because It'S Not Just What i Say On My Site Is The Truth And Nothing But The Truth So To Say Um, but looking at it from the buyer perspective.

00:20:11: It's also a window into even more information being compelled for me.

00:20:17: so how do you think about?

00:20:18: How that will change basically the expectations and also the behavior of buyers in their research journey?

00:20:25: Yeah

00:20:26: So I uh i think The biggest the biggest changer And i see Even In My Even In Myself Is That the expectation around speed and as I mentioned before depth of content is much higher now.

00:20:42: When you can go as deep in an LLM platform, that starts to outweigh conventional personalization channels historically would need to look at intent data or use a personalization engine for a certain degree.

00:21:00: understanding about a buyer then tailor their experience accordingly whereas They have the ability to do a lot of that for themselves.

00:21:08: And I think, To then land on a website where having done some of their research...to them and on site we're ultimately.

00:21:17: you know You've got very preformed preconceived kind of journey for them but your trying to map out.

00:21:25: um i think those two things can clash quite heavily especially when You know in the sass space for example if people want to.

00:21:35: See the product, If they wanted see things like pricing and what's a kind of get into that level of detail?

00:21:41: And thats all hidden.

00:21:42: then it is hidden behind.

00:21:43: typical can be to these ass products flow them.

00:21:47: i think that can become a challenge but even more broadly.

00:21:51: I think you have to kind now understand there are lot about by research.

00:21:58: is your perspective.

00:21:59: my research is done on in those LLMs before they even reach your website.

00:22:04: So being able to influence that as best you can means the type of content that you're creating needs to evolve, so I mentioned things like pricing there and being able make sure that pricing information potentially is displayed or with a service within these LLM platforms provided by a third party or by someone other than yourself, that all of the information you're giving all the way through the funnel is completely accurate and ideally your source of truth for all of our information.

00:22:38: And I think we kind have to accept that naturally we will start seeing a diminishing top-of-funnel where some of the traffic coming into the website certainly would've been in research or very early research stage historically probably not gonna come to your website anymore.

00:22:57: it's all going happen in the lm platform right and so naturally that means then you start into gear or your website up to be even more of a. Direct conversion engine in some cases looking at how can u overcome, When

00:23:18: I think about how basically the whole digital buying experience changes now, I'm often times reminded of how offline buying experiences changed.

00:23:28: Where people that have walked into a store ten years ago maybe looked around and went to someone who offers consulting there and said hey you know i want buy this or that do some recommendations whereas And these people that are a little bit tech savvy, they go heavily pre-informed.

00:23:46: They go to the shop and already know what they want.

00:23:49: It's just a matter of do you have it in stock?

00:23:52: Or like?

00:23:53: can you basically deliver it to me or whatever questions people have?

00:23:58: Do You also see this analogy from that point?

00:24:03: What do think we could learn?

00:24:04: maybe from retail?

00:24:05: because Retail has already experienced for quite some time Before it came to digital.

00:24:11: Yeah,

00:24:12: that's a great analogy and I absolutely believe we're at the inflection point again.

00:24:18: For me That inflection does everything at some point natively start To make sense in an LLM.

00:24:28: Do you get to a world where Eventually maybe We just all providing databases of information to plug into an LLM platform and everything is surfaced within that sort of experience.

00:24:42: I think even if you look, like Baidu for example in WeChat some of the platforms exist out of APAC.

00:24:53: generally they've been much more heavily personalised world where a lot their research happens.

00:25:03: I feel like we're more and more kind of heading towards that sort of reality where, yeah why does someone need to go to ten different shop windows or websites if they can do it all from one place?

00:25:16: You see rises in things like cloud co-working.

00:25:21: you know agentic experiences were people could effectively build out their own personal assistance.

00:25:31: Agent mode to do my food shopping for me and it is so.

00:25:36: why would I?

00:25:37: yeah, Why would i need to go to some of the websites in the future?

00:25:41: The thing that.

00:25:45: Is still important and users still needs.

00:25:49: they still like human component.

00:25:53: They want to be able to tangibly see this they want to buy, particularly when it comes to software and I'm speaking purely to the SaaS space at the moment.

00:26:05: But SaaS is complex for a lot of businesses, particularly larger organizations with complex requirements.

00:26:12: you can't necessarily just buy new software solution on the basis of a quick demo like ultimately... You need to dig into that in a lot more detail but i still believe LLN platforms natively, which to your point cuts out a lot of the high street browsing as it would have existed previously.

00:26:36: And do you see people actually booking a demo at Personio within the chat GPT interface?

00:26:45: So not just asking for what's the best HR platform etc.

00:26:50: then obviously Personio shows up And then clicking to the side or searching for person you and I'm going through this site, but actually completing that conversion.

00:26:59: Within chachi bt any time soon?

00:27:02: i mean who knows but ultimately why not?

00:27:06: yeah y'know why would they not do like it does happen?

00:27:10: um...i think there's a missing part of the gap.

00:27:16: their thing could be..they often wanna see before they try.

00:27:21: so You know that's not a long way away.

00:27:24: We're certainly not a million miles off, That level of kind of integration into LLMs like chat GPT etc.

00:27:32: so yeah I can see it being pretty viable.

00:27:36: you might even end up with things like marketplaces or what are built in to those LLMS directly or natively where people as they say feeding some information and then its getting served right?

00:27:50: I think it's a really interesting space at the minute.

00:27:53: There is no telling where these things will get to, but part of that barrier would be just native adoption of LLMs and i don't have good sense what current LLM adoption rates look like across different core parts.

00:28:16: our audience or certainly within kind of, you know B to be it's BSAS companies how heavily they're leaning into some of these things.

00:28:23: But that gap is only gonna narrow over the next twelve to eighteen months.

00:28:28: also I know like the vast majority people who speak to now do a lot of their early research with in those LLM platforms.

00:28:34: so hey i can fully see that becoming reality at someplace

00:28:39: and we should also feel used to be more like a performance or tech play, where you were looking for the latest hack and then the latest edge.

00:28:52: And how could maybe in a way trick the algorithm too now?

00:28:56: The whole space becoming more brand-aware.

00:29:00: so we have to think about How certain company is perceived across multiple sources.

00:29:05: We have to shape their narrative.

00:29:07: I feel just if i look at conversations that are happening around topic it's much less Or at least from, it's my personal perspective.

00:29:15: It's much less about.

00:29:17: there is this one technical thing and they're like the keyword density and this and that And then more like brand narrative in third party sites and review site on all of them.

00:29:26: So what your perspective?

00:29:28: Yeah its almost kind of blended a few different worlds That perhaps would have been more disconnected previously.

00:29:36: Right.

00:29:36: so to you point I think Brand Perception Reputation Management But also conventional kind of SEO and content strategy are all overlapping a lot more heavily than they did previously.

00:29:49: I think naturally, there is still hacks out there in their still things that happening to try and kinda hack the algorithm because it's still changing rapidly but... Naturally you know i've seen alot of those.

00:30:07: even if immediate spikes as a result of some of those hacks, generally they soon get leveled out again longer term.

00:30:16: So I think being able to more heavily now align across areas like brand conventional SEO, AEO and tech SEO optimization is still super important.

00:30:30: just maybe we've probably dialed up in slightly different areas there.

00:30:35: so previously it would have been making sure that tech health was in order.

00:30:42: Now it's more, okay we're using the right structured data.

00:30:45: is schema markup being used effectively?

00:30:46: to your earlier point should we be testing?

00:30:49: Markdown specifically for LLMs and specific crawl looking at crawl past that LLM are taking across the website some of those sort stuff.

00:30:58: so I don't think anything that anyone really has been doing historically of an up-waiting.

00:31:08: now into content delivery and strategy.

00:31:13: And as I say, that offsite content strategy has probably evolved more than anything else at the moment because it's so important.

00:31:22: We obviously touched already a lot on AI as channel especially like AI search.

00:31:28: but you also mentioned AI for example, where you do grocery shopping with Agent Mode which is obviously something that makes a lot of sense.

00:31:39: Sometimes when I get home from work and want to make nice dinner This would be something that AI definitely helps with.

00:31:56: Unfortunately, it's still hard in the physical world but anyway we're drifting off.

00:32:02: One of the obvious use cases also in marketing when it comes to using AIS a tool is content generation.

00:32:08: this Is Something That I think Also Was one Of The earliest Things.

00:32:11: We Used It For Either Drafting This Like Campaign Email Campaign a newsletter thing, we may be used for drafting like a poem or something.

00:32:22: Some things work better some things worked not so well.

00:32:25: So I'd like to get your take on AI generated content?

00:32:30: Do you think it's basically Something that We shouldn't lean into You shouldn't use?

00:32:36: Or do you think It is definitely That has To Be part of the future marketing?

00:32:46: Ultimately, I do believe it should ultimately be part of the future over marketing.

00:32:51: It comes with a heavy caveat that still has to follow all conventional rules around good content creation and needs to be authentic or trustworthy the right signals for it to be valuable.

00:33:12: For a real person as well, there's an LLM and I think you can get through that point with AI but a lot of depends on how well trained your AI is?

00:33:23: And ultimately some people will lean into just jumping straight in chat GPT blog post or an article, or a listicle of whatever it may be.

00:33:35: And they fed in very little source material so just churns out and regurgitates the stuff that already exists on market.

00:33:42: if you're plugging into your own data sources conversations with any learning from customers potentially to case studies research brand tone of voice and guidelines like previous examples have content that you've created over time, it will become much more intelligent.

00:34:07: And the quality about pool becomes stronger as well.

00:34:11: I think thats where is really helps with speed execution particularly in world of AEO.

00:34:16: were wanna be able to spin up tweaks and changes to content that previously would have had.

00:34:23: So we've gone through some, some crazy workflow like ideally you should be in a position where you can identify what's on of those areas should be the content gaps?

00:34:32: And like within twenty minutes it is built on your website and its updated and live right Like That's The Future.

00:34:41: But It has To Still Be Relevant And Of High Quality.

00:34:46: I think there Is A Lot Of Stuff Out There At The Minute Doesn't doesn't solve for that gap, you know typically see the any of that kind of slot content That gets spun up very quickly may have an immediate Ranking benefit and or may having a media benefit in terms of kind of citation in in LLMs But then that that soon drops off afterwards like I've seen numerous examples with where this happened.

00:35:10: And why do you think that is?

00:35:11: So many people still Do The do the content with AI.

00:35:16: It's not fat with a lot of context and not fit with actual customer conversations, everything you just mentioned.

00:35:24: You described exactly how we should do it?

00:35:28: And still I see people when they check sites or look at the content that is actually on these sides... ...I can really smell this slop!

00:35:38: So why does it happen?

00:35:40: Because models available to all us.

00:35:43: They are so powerful now, GPT-Five point three five point four Claude Sonnet.

00:35:49: Four point six.

00:35:50: Opus four point six.

00:35:51: these were super powerful models that everybody can access basically with a ten euro or twenty Euro subscription a month.

00:35:56: So it's not the financial issue obviously

00:36:00: No I think there is probably numerous reasons for.

00:36:04: but in some cases It could be people relatively early on that AI maturity curve generally wanting to wanting to get involved and wanting to try something new, test something but not necessarily having all of the broader context around.

00:36:20: what does that actually mean?

00:36:21: What good content versus kind AI slop look like.

00:36:26: Maybe they're putting too much trust in what the AI is coming back with then not realizing it's probably built off ten of the same bits of content already exist out there at market right now.

00:36:39: so I think the level of integration that you can do within some these tools is organization dependent.

00:36:48: There's a lot companies I speak to and even very large, very tech focused organizations... I would say laggards in terms of AI adoption ultimately.

00:37:01: so they don't have integrations or data sources.

00:37:05: there are more trust.

00:37:09: issue around AI generally, so naturally what they're feeding it is not going to create as powerful a model for them.

00:37:21: So I think those are couple of areas and then the other might just be Feeling the pressure in the urgency of.

00:37:35: you know we need to do a i stuff and that's probably something is echoed within their companies as well.

00:37:40: as there are organizations looking around.

00:37:43: other company said talking about ai.

00:37:45: You see the echo chamber on linkedin.

00:37:47: everybody now that isn't an expert.

00:37:49: so does naturally this kind of almost unspoken peer pressure of everybody looking at, do some stuff and show that we're using AI.

00:38:03: So I'm going to create some AI articles, then get them out there as quickly as possible and see what

00:38:07: happens.".

00:38:08: That naturally is still happening.

00:38:11: And the more mature companies when it comes to AI usage and adoption have probably they've already been through that learning experience and come at other sides are now starting to refine their content models whereas there's something in that journey of the minute for themselves, ultimately.

00:38:31: And what would you say?

00:38:32: How can people build more resilience against that outside pressure?

00:38:36: because I talked to like from junior to head of two director to sea level and nobody is immune to that.

00:38:45: so they're individual People.

00:38:48: some are more somewhere less.

00:38:49: but What Would You Say?

00:38:50: Because Obviously Being Just Driven By Every Next LinkedIn Post Is Not A Good Strategy.

00:38:57: But Still I think a lot of people can relate to that.

00:39:01: Damn, maybe i should look into co-work more?

00:39:04: Maybe i should build something with cloud code... ...maybe i should do that.

00:39:07: but then you look at your schedule and you only see so much time being free!

00:39:12: So what can you share also from personal experience on.. ..what people could do around building more resilience?

00:39:19: Yeah i think it comes down to In some cases that the individual right but also just generally speaking i would say having having curiosity so you know wanting to kind of go out there and understand.

00:39:38: This of their own volition before it gets to a point where there's a link in post that pops up and they're like, uh you know damnit.

00:39:43: I need said i need to do that or before the c-suite?

00:39:45: will somebody else sees this?

00:39:47: is that linked in post?

00:39:48: i think he was important.

00:39:50: i think if your learning new experimenting for yourself then u can even get ahead over lot ove alot those things and i think truly thing about what problems are that you actually want to solve with ai as opposed to what is the ai That we wanna try and get to solve a problem?

00:40:05: if that makes sense like i think ultimately you will know what u should be able to kind of no within your, automate and reduce a lot of the manual work or pressure, et cetera.

00:40:29: And I think that's where then you can start to do more of that learning and then take the learning in.

00:40:35: you can build on that ultimately.

00:40:37: but when it comes to broader pressures trying be ahead of that curve spend time experimenting even just your personal life realize some value.

00:40:53: Yeah, try not to get too distracted by the noise that's out there.

00:40:56: Just focus on what.

00:40:58: are they key problems you actually want to solve using AR?

00:41:01: And how do you think about like marketing org of future because I see more and more conversations going around Do we still need junior roles Because we can use AI for that We can deploy agents etc.

00:41:16: And I mean, obviously also in your role looking at your team.

00:41:20: Maybe you're also looking other teams and personia or the teams that you talk to are.

00:41:27: what would you say?

00:41:28: What still holds true for the future about what we learned... ...what makes a strong marketing arc.. ..and what do you think is one of the biggest changes happening currently?

00:41:37: Yeah!

00:41:39: kind of very high level.

00:41:41: without going into too many specifics.

00:41:43: I think ultimately for me it would be that, its just a change or an evolution in terms.

00:41:50: everybody's role and what i mean by that is the natural marketing has gone through this shift over past ten years more.

00:42:00: anyway where we've seen rises in also we see falls areas were Previously, you know everybody was more of a T-shaped marketer.

00:42:11: I remember when i first started out.

00:42:15: the digital markets at that time were responsible for everything from website optimization through to performance media campaign builds like whole host if different things and then naturally over time A lot of roles have become much more specialized.

00:42:30: And I think now what we found is that AI helps to smooth out some of that experience layer a little bit.

00:42:40: To the point where people are starting become more T-shaped again, suddenly you've got... People who had great ideas and were able to prototype or bring those ideas into life.

00:42:50: You have people with content writers but now can be more content writer.

00:42:54: People previously weren't analytical so they're much more analytical because they use AI to help them.

00:43:00: So I don't see there necessarily being shift.

00:43:04: Do we need to remove junior roles or do you need certain roles?

00:43:07: I think it's probably more a case of just everybody is on, again that maturity curve of AI adoption in the level of T-shaped marketer they're able be.

00:43:20: Maybe end up with slightly less silos and typical subteams within marketing.

00:43:25: but actually perhaps everyone has project particular area of focus aligns, but actually they can own a much broader email scope within their own world.

00:43:37: Who knows not sure just yet.

00:43:40: and do you think everybody has to become a coder with codex or Claude Cote?

00:43:46: I don't think anybody has the beam.

00:43:48: i think everyone should be.

00:43:51: Yeah

00:43:52: even content writer

00:43:53: yeah i think.

00:43:54: so i think i think every body can really kind benefit from building and shipping in learning and kind of understanding what that's like.

00:44:06: And again, not even necessarily a business context in their own personal life but being able to bring something to life the existing your mind you wanna build tangibly is super rewarding.

00:44:19: there are lot people who have been able do it previously.

00:44:24: so going through Going through that experience and building something, just kind of learning more about the process.

00:44:34: Even using it as a tool to bring any idea you have in life... The amount of times I've used cloud code or lovable or whatever else to articulate vision around something Or skip building a slide deck and build an interactive report or dashboard.

00:44:54: But God said it's so helpful.

00:44:59: It doesn't mean that everything you build has to be something taken and shipped directly, like sometimes is just a better way of visualising something or bringing your idea into life.

00:45:11: We even have been playing around with shipping for prototyping much more heavily in code as opposed doing all the prototyping in Figma which helps us jump straight.

00:45:26: Okay, so what are the actual functional requirements?

00:45:31: What's the behavior like?

00:45:33: how does that feel or just experience feels outside of having to build a lot of stuff in Figma.

00:45:37: So I think there is a lot use cases across every department and role it could benefit from.

00:45:43: coding more broadly.

00:45:45: Really loveable has this mission?

00:45:50: basically everybody can be your creator And like AI is your creativity enabler, so to say.

00:45:56: Would you agree with that?

00:45:57: Yeah I would.

00:45:58: yeah definitely.

00:45:59: um i mean you obviously need

00:46:01: the...

00:46:02: You still need the ideas!

00:46:04: You still needs be able kind of think what's this thing that want me build?

00:46:09: and none other than Any of the tools out there are truly production ready in any capacity just yet.

00:46:18: I think that will come eventually, but i don't know they're at the moment.

00:46:20: But it does mean yeah everybody can be a builder.

00:46:23: Everybody comes up with something or bring their idea to life.

00:46:29: It's so common you have conversation with a friend or colleague and they say i've got this great idea that I wish I'd shipped so.

00:46:39: I thought of the idea been done, which I created because somebody else has done it now and that's where this company is come from.

00:46:46: end doesn't solve for all of them but certainly help people take step in right

00:46:50: direction.

00:46:51: okay thats very interesting and like to get your thoughts on something that I have been pondering little bit over last weeks month because I feel like AI and creativity, a lot of people are skeptical about that.

00:47:04: two words in the same sentence.

00:47:06: But i actually felt there were lots of people from the corporate world so to say have been... So we tell them how they think outside-of-the box but don't really want us to do their job as quickly as possible.

00:47:23: Now when you look at my own team at Radiant they experienced what AI enables them to do and exactly to your point, you have an idea.

00:47:35: And you see that you can bring it to life the more you train that muscle of creating ideas off not limiting Your creativity?

00:47:45: I had this same experience very early in my career when i was building websites for people... ...I always felt like okay.. ..I cannot propose something that I cannot build and I limited myself heavily.

00:47:57: And then at some point, it was just... It was twenty-fourteen.

00:48:00: so way before i made the switch!

00:48:03: I felt like you cannot limit yourself to what your able build today.

00:48:07: You have to propose whats best Then figure out how to build afterwards.

00:48:12: So do feel that makes sense in any way?

00:48:16: What I said

00:48:18: is

00:48:18: something where people need to built muscle.

00:48:21: We had to build visionary muscles a little bit more again.

00:48:24: Yeah, it definitely resonates because I think there's probably a lot of people out that thing about.

00:48:31: they have those ideas.

00:48:33: Be for campaign landing page or brand campaign whatever else might be like.

00:48:39: this is always a lot.

00:48:41: creativity exists but the barriers pop up are probably.

00:48:46: is that gonna be too much work for the team?

00:48:49: Is it going to cost you money, or require extra budget.

00:48:52: Is there something they don't have capacity support on?

00:48:56: therefore what's the point of suggesting I'm not gonna bother and stays in their notebook.

00:49:01: so i think thats thing this helps overcome.

00:49:06: we see some great examples even within personia where people who've had these ideas been able to jump into core code or prototype something up and share it out across the team.

00:49:21: And actually people are like, damn yeah that's great!

00:49:24: Yeah let us build this.

00:49:25: Let us focus on doing

00:49:25: it.".

00:49:26: I think that is really powerful about these tools.

00:49:32: Don't get me wrong... It can potentially become chaotic in future if you end up in a world where suddenly everybody at the company is prototyping stuff they want to build but still don't.

00:49:45: Those ideas flying around is a bad thing as long you can actually operationalize it and bring them to life, which often the blocker.

00:49:57: It helps build that creative muscle in people who start thinking more about what's next evolution?

00:50:13: Which particularly in the world of kind of marketing and where been digital is, something that everybody enjoys doing and ultimately is why everybody got into the business in the first place.

00:50:23: Yeah, I don't want to switch seats with a head of engineering because i cannot imagine the amount of commits and pull requests etc.

00:50:32: on github That you get if Everybody In The Engineering or goes To Cloud Code Or Codex And They Can Just Like Bump Out Ten X The Code In The Same Time Even We If we Do A Little Bit Of like Internal Tooling And We Have To Coordinate Between True Three People Obviously we're not engineers, but this already feels like on the edge of being genius and being madness.

00:50:57: So

00:50:58: yeah... Yeah I think the exciting prospect there is that.

00:51:03: so as it stands at the minute i agree.

00:51:07: That probably causes some pain particularly if a lot quality code isn't necessarily up to standard.

00:51:14: you've got considerations around kind of security and broader integration into tech stack, things like that.

00:51:20: But again

00:51:22: if you just

00:51:23: look forward to the not too distant future in being able to make sure there's nothing to stop building an agent is trained on your code base integrated into a web CMS or whatever else such that You can give people the autonomy code base ultimately and your own integrations input.

00:51:50: so you move away from someone building something in loveable and sending it to your engineering team game.

00:51:55: please can you build this thing for me.

00:51:57: To them having to then completely rebuild a different scratch to two ultimately some someone that can, build up or certainly kind of get something very heavily heavily the right direction if they want to build exists based on the code base already is there.

00:52:16: so The thing is, again it's an exciting time.

00:52:20: Let us tap into one aspect of SEO AI search before we wrap up the conversation a little bit which basically how to balance investment being at money-time people on team.

00:52:36: between these two like between SEO and between AI Search.

00:52:40: I mean even uh, should we keep doing what we did for SEO?

00:52:46: Should be completely lean into a search et cetera.

00:52:48: So how do you balance that in your team at Personio?

00:52:53: Yeah I think the-the-the balance for us really is all of it.

00:52:59: The common layer that spans both those things Is content and good quality content right.

00:53:05: so That's where its important to be making some of that investment.

00:53:14: SEO strategy, AEO strategy both very heavily aligned and the tooling then is deployed to monitor across those different things so you can make sure they're right decisions equally important.

00:53:30: but all of this falls over if we don't have people who actually write or create content.

00:53:35: To our earlier point, yes.

00:53:36: You can do some of that using AI but you still need to bake in things like deep localization and get supermarket specific work very closely with obviously your regional teams in certainly our instance But your product marketers Your subject matter experts internally.

00:54:00: So having strong content creators and coordinators that can help with that is kind of the key area.

00:54:08: And certainly where we're, as I say investing some at that time.

00:54:12: And then AI wraps around all of that as well.

00:54:16: it's other areas so were very focused on.

00:54:18: how can we streamline those content operations using AI much possible?

00:54:24: How could potentially explore bringing that AI content generation to to life in a way where it is trained.

00:54:34: It's kind of fully ramped on all the different content data points that we have or sources, so that we can build and ship them into hours not days weeks or months generally speaking but contents at heart for sure.

00:54:52: And if you look at things have entered the conversation, so to say around AI search.

00:54:58: So one thing you mentioned earlier also Reddit red community engagement if you want to call it like that.

00:55:06: and then on the other hand something I mean there are lots of things but just give two examples.

00:55:11: thinking about a third party mentions more news outlets mention new stories which can basically push their what do think temporarily hyped and things that have come to stay.

00:55:27: Good

00:55:28: question, yeah.

00:55:28: so I would say there's been a lot of hype around as you know and just mentioned certainly some community sites like Reddit we saw very early on everybody suddenly downed.

00:55:44: tools in their content strategy evolved into how can we spam the comment threads?

00:55:49: I do think that has, thats already peaked and yes you still see reddit sighted pretty readily.

00:55:57: but i would say some of the community management aspect is probably biggest hype there.

00:56:05: Is now kind shifted into more conventional onsite content strategy work like im seeing your own website if it's created well as a bigger influence in a big bearing than some of those some of those third parties now, depending on the channel.

00:56:23: Then Google still leans quite heavily into Reddit particularly with AI overviews.

00:56:27: but certainly a lot of chat GPTs and stuff like lean more in to first party sources.

00:56:32: I've found.

00:56:34: so i would say that aspect is probably one that's less immediately critical although we are working making sure areas social channels get adequate engagement.

00:56:50: And I would say areas like review sites and third party websites are still very, very much important and key.

00:56:58: There's two approaches to that.

00:56:59: one is you can encounter some of the content created there in use within your own website or go and obviously influence that content as it exists on a third-party site.

00:57:13: but those

00:57:17: And based on that, would you also say that improving and maybe adjusting the strategy around the content you have in your own site?

00:57:28: That this was at least until now the biggest lever you've pulled for your AI search visibility.

00:57:36: Yeah I still would say it's the biggest leaver but i definitely find our website has the biggest bearing visibility within our particular industry.

00:57:48: that may vary across different industries, but for us it's the biggest lever.

00:57:53: That is what I meant about being able to counter some of the content that exists out there at the minute and an example might be a third-party listicle website talking perfect solution for small businesses in Germany only, and to counter that we're able create content which actually talks about our presence on other markets.

00:58:20: It's the scale of what we operate with a large number of customers within the mid-market enterprise as well just like an example.

00:58:31: so those sorts of things then mean naturally you can count some of this preconception that the LLMs have.

00:58:40: But that said, you know the other angle is you'd go and partner with whoever.

00:58:44: The owner of that particular list to call us in your get it updated at that source as well.

00:58:48: then you've got two places where there's being referenced.

00:58:52: so thats kind a typical approach.

00:58:55: now we always want make this podcast as actual possible.

00:58:59: I think were already delivered on their promise.

00:59:01: but if fellow marketing leader came too you or listen to this podcast right now and said, hey we need to get serious all about AI search.

00:59:12: I feel like we are missing out there.

00:59:14: what would you tell them to do first?

00:59:17: I would actually say blockout depending on the level of research or work that you've already done.

00:59:25: even if you got no tooling You have no AI search monitoring tools as it stands.

00:59:31: Blockout half a day Or couple hours.

00:59:36: spend time reverse prompting within some of those lm platforms to understand how you're being perceived.

00:59:42: so why i mean by that is pretend ure a. Pretend your one of your customers and go in perform the number of different searches.

00:59:49: if things do you think your customer would be searching for, and see what shows up.

00:59:54: and when you see what show us ask why did you choose this?

00:59:58: why did it use those things like why did he pick those companies as listings?

01:00:02: and then if you are not on list ask why?

01:00:04: Why is X company, why was my company not in that list?

01:00:08: And it will give you some really kind of insightful information.

01:00:11: You can even then ask what did you use as the cited sources for that information and that gives an immediate sense than if... What are those offsite or onsite bits of content maybe doing you harm or causing your own issue?

01:00:29: from there you could do some actionable next steps without needing to go invest in a LLM platform.

01:00:34: Nice, I think that's definitely some solid piece of advice.

01:00:39: Now i have a final question and then always ask that I bluntly stole from Lenny's podcast who is the significantly better podcast host than I am.

01:00:50: so if ever come close to him... ...I can retire.

01:00:55: but my questions are what didn't we talk about?

01:01:03: That's a good question.

01:01:05: Thanks to Lenny,

01:01:07: thank you Lenny.

01:01:08: I would say that the thing we probably could talk about more is or should have talked about it a little bit more.

01:01:17: maybe measurement of all this stuff right?

01:01:19: So i think measurements are.

01:01:23: a lot of marketers get asked around LLM visibility in particular and AEO at the moment Probably more so than some conventional SEO SCO metrics and stats as it stands at the moment, but there's obviously a lot of different ways that you can measure AI and LLM visibility.

01:01:39: Is is it?

01:01:40: The number of citations...is your share or voice...?

01:01:42: Your overall visibility score or ranking..?

01:01:46: Is this position in which you appear in the prompt result like.. There are other things here.

01:01:50: I don't think anyone yet has truly locked on what is one metric we care about?

01:01:56: Because it sure as hell isn't going to be traffic coming from those channels.

01:02:01: I think there's a big shift that is happening, we still see decreases in like across every industry and more of the traffic coming through conventional organic traffic coming though other routes.

01:02:19: now you see declines in PPC traffic volume as AI overviews eat away at top page on Google.

01:02:26: so metrics, measurement tracking reporting in this particular space is super interesting I would say.

01:02:33: And can you share something about maybe things that you tried or things that do believe for Sonio like where are at with measurements?

01:02:44: Obviously some stuff doesn't hold anymore.

01:02:46: but what's still viable and should people think more?

01:02:52: Is it self-reported attribution?

01:02:53: Yeah, so there's

01:02:58: a few different things but I would say naturally carrying less about traffic generally.

01:03:06: That is the big adjustment that we've made.

01:03:09: it like traffic... Is bit of vanity metric.

01:03:12: We pivoted awhile ago to think more about high intent traffic.

01:03:15: So you know people who have that intent want to buy or convert.

01:03:21: And then more generally, you know we still care about obviously things like keyword rankings and visibility in what?

01:03:27: not from an SEO standpoint but.

01:03:29: More generally just looking at are we influencing the metrics that we care about which is ultimately a wee getting kind of MQLs and pipeline et cetera for their business?

01:03:42: That's stuff that we carry out with.

01:03:45: channel attribution has become slightly more fuzzy You know, naturally you would see that people would perhaps go from an LLM platform and then either come direct to the website or maybe they Google through a branded term and comes through organic.

01:04:03: Or they Googled through a Google branded term in comes repaid.

01:04:05: so I think naturally like there's more of that kind of fuzzy layer around channel specific attribution at the minute.

01:04:13: um yeah may be y'know At some point we'll look attribution or better ways to track that.

01:04:22: But for now, we're more focused on.

01:04:24: yeah are we ultimately kind of maximizing that share of voice and are we driving the conversion that we want

01:04:30: in?

01:04:30: Interesting!

01:04:31: James thanks so much for sharing with us all those other thoughts like also background stories.

01:04:37: really interesting what you shared.

01:04:41: if people wanted What's the best place to follow?

01:04:51: I would say probably LinkedIn, like bio means if people are interested or want more connect with me on LinkedIn.

01:04:56: Occasionally less than i should but occasionally a stick post out some of this sort stuff as well.

01:05:01: so bio means feel free to follow me there.

01:05:06: Nice okay then thanks so much.

01:05:08: guys and girls go to James Ward on LinkedIn.

01:05:12: we will obviously put link in description for podcast.

01:05:17: Connect with him, please don't spam him.

01:05:20: Only connect with generally interesting questions or something like that.

01:05:25: and James thanks so much for taking the time today.

01:05:28: it was a pleasure And hopefully we'll catch up soon

01:05:32: Likewise!

01:05:32: A pleasure.

01:05:33: Thank you Nethus.

01:05:34: Bye bye

01:05:35: James

01:05:35: See ya.

New comment

Your name or nickname, will be shown publicly
At least 10 characters long
By submitting your comment you agree that the content of the field "Name or nickname" will be stored and shown publicly next to your comment. Using your real name is optional.