From “words for money” to rethinking the value creation process

Noz Urbina: Hello everybody and welcome back to the Omnichannel X Podcast. I hope you had a good break this summer. It is September 2nd at the point we’re recording this and I am here with Bouke Vlierhuis.

Noz Urbina: So if you don’t know who I am, I am Noz Urbina, the host of the Omnichannel X Podcast. And Bouke is an old friend of the podcast. He was one of the first speakers at the physical Omnichannel X Conference. So he was actually on the stage with us when we ran a physical conference in 2019.

Bouke Vlierhuis: Legendary. Still my favorite conference I ever spoke at.

Noz Urbina: Yeah, it was pretty amazing. And if anyone ever did go to those conferences, they were very special online and offline. The conference market is a little bit of an interesting thing post COVID, so we are happy to be a podcast now and still be friends with some of the connections we made from back in the day.

Bouke is a B2B marketer and early adopter in AI. Works a lot in content creation, but with a very strategic bent and a little bit of a technological know-how as well. Bouke, do you want to introduce yourself and tell the people a little bit about your background?

Bouke Vlierhuis: Yeah, so I recently did a podcast called The Origin Story Podcast, and I got to talk about my background for an hour. So stop me if I go long now, actually. I’m an engineer and once you train as an engineer, you’re an engineer for life. So I was trained as an aerospace engineer.

I never worked in aerospace. I worked in IT software engineering for five years or so. And an engineer to me is someone that gets a problem on their desk, then starts gathering information and once they have enough information, they come up with a solution. And the first solution never works. So they come up with a second solution and a third one, and they keep going and going and grinding at it until it works. That’s what an engineer does.

I tried to apply that way of working to writing a novel and that didn’t work. So I started freelancing just to get away from my novel, and that was in 2003, I guess. And I started building websites and I found out that people were really bad at copywriting for some reason, and copywriting just came natural to me. So I started out in politics, like ghostwriting columns for local politicians and that stuff. And all the while I was programming and developing like the things that you now get for $10 a month SaaS for that didn’t exist back then, like expense declarations or time sheets and that stuff. So that’s what I built.

And then at some point in 2019, I decided now I’m going to just do marketing for IT companies. Because apparently being an IT engineer still just made me very good at talking to other software people and that was really the moment that things started to take off.

Noz Urbina: I think that makes sense for me because often in content creation, you’re trying to achieve empathy. So if you do have personal domain experience and you can write, which is more of a, I don’t know how, I’ve taken writing classes, but I am of the belief that certain people can write and certain people can’t. It’s like you can teach dancing, but you can’t teach great dancing.

Bouke Vlierhuis: You couldn’t teach me dancing though.

Noz Urbina: So you just happen to be a writer who trained as an engineer and found your way there.

Bouke Vlierhuis: Yeah, yeah. Well, I think it all starts, and this is more relevant to the whole AI story than you might think. I think it all starts with reading. So our brains are basically wired for pattern recognition. So the more great texts you read, the better your brain will recognize whether or not you’re writing something good.

And it’s that feedback loop that actually makes writing good. So the thing I do is I read a lot of nonfiction, like piles and piles and piles and piles. I still do like 30 to 60 books a year. And that’s what people forget that they want to learn how to write, but you know, do you know how to read?

Noz Urbina: And how to read? I know many great readers who’ve lost the ability. I think that’s something that’s harder and harder to say, like, I will sit down and read a book in our current age.

Bouke Vlierhuis: Yeah. Well, the great thing about this home office that I have is that this is my office and my home is back there, and when I stop working, like at the end of the afternoon, I leave my phone here in the office, and sometimes that lasts for two minutes. Sometimes it lasts for the entire night. But it gives me a window to do like the analog things like cook dinner, talk to my kids and read a book.

So if your phone is within arm’s length, forget about the book. I mean, my brain is just as addicted as everyone else’s. But you leave your phone at the office and then you read your book. If you take one thing away from this podcast, this should be the one.

Noz Urbina: Okay. Well, speaking of books and feedback loops that you get from creating content, you just are now a published author. I got my copy last week.

Bouke Vlierhuis: Yeah, yeah, yeah. I’m extremely proud to be in your bookcase.

Noz Urbina: And if only I could read Dutch. But I did read the Google translation and I really liked it. There was actually a lot of nice, fresh concepts in there, which is rare. So the survival guide for the B2B marketer, AI Survival Guide. Best translation of the title. You want to tell us a little bit about why write this book?

Bouke Vlierhuis: Well, so AI happened and it started taking sizeable bites out of my business model quite quickly. Like March 2023, I was seeing that the first clients leave. So that’s pretty soon after ChatGPT landed because I was working with IT companies, still do, and they’re quite quick adopting this sort of technology because they were already writing their code with GitHub Copilot.

So they were used to using these AI tools and they were pretty quick in going like, “Yeah, I ordered three of these web pages. But I already sort of made them myself. Can you look them over?” And I was like, oh no, this is not happening. I am not gonna be your look over the ChatGPT texts guy. That’s not gonna be me. So I was like, okay, I need a strategy.

And then I was thinking about my own strategy and then I realized, okay, my clients need a strategy as well because this is coming for us. I mean, this is coming for our entire process, the entire way of doing things where you get to sit down and think of a strategy and think of certain content pieces, like, we’re gonna do four white papers this year, and that’s ambitious, right?

And then you brief your white paper and you wait for two weeks because I’m typing, you know, I spend my days typing. I make my money typing, and then I send it to you and you look at it and we do feedback and I gotta go back to typing. And that entire process just sort of got smashed into like nothing.

And I was trying to deal with that, being a copywriter. But my clients were also trying to deal with that, and they were asking me, so how do we deal with this? So I was just, I was like, okay, so I can’t make money typing anymore. That was just done within the year basically where I was working, like in the IT sector, like the whole thing where you can sit down for a week and write a white paper that was gone. That was so quickly, you just blink your eyes and your entire business model is gone.

But my clients were asking me, what do I do with AI? Where do I deploy this technology? And I started advising them. And that started the book basically. So the first idea was, okay, so my clients are asking me this. And if your clients are asking you questions, that’s a marketing opportunity. That’s how marketing works. Client questions. You answer them in the right way. You get new clients.

Noz Urbina: You’re preaching the choir right there. Questions.

Bouke Vlierhuis: You should probably write this down.

Noz Urbina: Exactly. That’s a good one. No, questions over time. For those unaware, Bouke has just touched on one of my hottest hot buttons about how all design, content design, user experience design, et cetera, is the mapping of user questions over time. That’s what we’re here for, is to help people find out how to answer their questions in a way which leads them to their objectives. So, absolutely. So that is marketing. It is design. It is designing, making the customer experience.

Bouke Vlierhuis: So I’m a B2B marketer, and I provide B2B services myself. So in B2B marketing, whatever the question is, the answer is always an ebook, right? So I started writing an ebook about how B2B marketers need to deal with AI and that spiraled out of control basically, once it hit like 30,000 words. I was like, okay, maybe this is not fit for a PDF.

And then I thought, okay, I’m gonna make a physical book and I’m gonna just pass them out as a present, like as a business card. And then I onboarded my favorite designer and I asked her to make a great design for it. And the moment I got the design, it was like, okay, this is way too beautiful to just be passing out at corporate drinks or networking events.

So I took a deep breath and called one of my favorite self-publishing houses. They did my poetry collection last year. So I consulted with them and then I decided to just go all the way. And now I got an email this morning that some person in some bookstore in a town in the Netherlands where I’ve never been bought my book. So that’s magic. It’s magic. So I went all the way and it’s now on all the online bookstores and it’s even selling. It’s quite niche actually.

Noz Urbina: Let’s focus on a little bit more of the meat in the book. So you and I are both big on process. I think that’s something that we both have a somewhat of a technical background moved into strategy, content experience. And I think AI is kind of bringing the nerd into art class.

We have been advocates for years telling marketing teams that they needed to have process, they needed to have structure, they needed to have defined life cycles and knowledge pipelines, information pipelines, data pipelines, governance, all these fancy things to basically say, get your house in order. You need to have a repeatable model and system to scale content. It is a creative artifact. There’s human creativity involved, you’re setting up a whole factory.

To scale requires you to set up a whole factory with a supply chain and partnerships, inputs and outputs. So if you’re in B2B marketing, it’s not just a bunch of artists creating as they want to. So how does that, you have AI, which is a technical tool, which is transforming that business. How do you bring that into the B2B marketing world, and then what are the big effects?

Bouke Vlierhuis: Well, yeah, I’m a process analyst by background. I did some process analysis quite recently actually in a lab. And when you start looking at processes, there’s always one, the layman’s term is the bottleneck, but if you’re into processing, it’s the constraint, right? There’s one constraint. And anything you optimize outside that constraint is meaningless because the constraint will always be the constraint.

And classically, in content creation, creation was always the bottleneck, the constraint. And that constraint was so huge that it was pretty much useless for any B2B marketer to be looking at their process because whatever they would optimize on their side would be consumed by me, the copywriter or the designer or whatever, taking two weeks to write the thing.

So the two weeks to write the thing turned into an afternoon. And then all the other constraints and all the other bottlenecks in the process, they start sticking out. So that’s what happens when you optimize one part of a process. The inefficiencies in the rest of the process come out.

So us marketers, we’ve always known that approval and briefing processes were extremely inefficient. We’ve always known this, but nobody really cared about optimizing these things because creation took so much time and that just changed.

Noz Urbina: And put out there, you know, whatever error.

Bouke Vlierhuis: Yeah, exactly. You took two weeks to write the thing so I can take two weeks to read it, right? That’s how it works. And now, another thing has happened and that content has become smaller and faster. So we’re moving away from white papers and we’re moving towards like thought leadership on LinkedIn, or we’re doing like video snippets or live streams or whatever.

So content production has had to scale up and get faster. Without AI even. So these two things have combined and suddenly all of a sudden, like process analysis is hip again. I was at a marketing event and the first speaker just pulled up his first slide and it was a flow chart of their marketing process and I sort of stood up and went, “Yes! Flow chart! Finally!” 20 years in marketing, and this is my first fucking flow chart.

So yeah, process analysis has had a complete renaissance in marketing and I love it. And that’s the meat of the book is about, you know, not about the one thing I didn’t want to do was make the next like prompt book, so open up GPT and type this, that would be outdated in two weeks, right? When GPT-4 replaced GPT 3.5, I got to throw away all my prompts and start over so that happens like every couple of weeks.

I wanted to teach marketers something that they’re not very good at, which is have an overall vision of what technology means for them and have an overall vision of their process, starting with what is it that we’re trying to achieve here? Because marketing is very, they often behave like an internal agency where sales or a director or whatever will come. We need a white paper and they start running around and trying to create that white paper.

So I think AI will eventually help B2B marketers take that step towards being more strategic, being at the table when the decisions are taken, knowing that creation will not bog them down. So knowing that they can be strategic, take the time to talk about strategy and to talk about purpose and all these things that, well, we’ve used the words, but eventually, when you get back to your desk, it’s just trying to get the thing made right. And AI will help them do that. And I tried to write a book just giving them some pointers about the new mindset that you need.

Noz Urbina: You had some terms and concepts which I really liked. So for example, the chainsaw effect, which sounds like what you’ve been talking about.

Bouke Vlierhuis: Yeah, absolutely. That’s exactly what I’ve been talking about.

Noz Urbina: For me that’s if the way that I understood it is like you’ve gone from the axe to the chainsaw. And it used to take you so long to get your hands on a tree that everything about processing the tree and distributing it and products and things was not the problem because cutting the thing down took so bloody long and then someone hands you a chainsaw and suddenly you’re exposing all of the downstream issues in your process.

Bouke Vlierhuis: Absolutely. Yeah. And this comes directly from my experience gardening, where I love electric tools, but eventually, you cut down one tree and that takes you about like 15 minutes and then you look behind you and there’s so many branches and stuff that takes you the rest of the weekend to clean up. So that’s where the chainsaw effect comes from.

Noz Urbina: I love it when somebody, I think that it’s always great and I find a lot of great thinkers that they’re crossing the streams between something they know in their life. It becomes a kind of a metaphor to help them understand a new concept. The other metaphor I really liked, which we can discuss the origins of, was the fountain pen syndrome. Do you want to tell us about that?

Bouke Vlierhuis: Yeah. Yeah, that actually comes from a Noz Urbina quote. So I think we did a two part podcast like a long time ago, and then you said something that resonated with me, which is people identify with the inefficiencies of their trade. So there’s copywriters that identify with the click clack of their keyboard.

And I started thinking about that and then I started getting really annoyed by the logos that copywriters and creative agencies have. Because they always have like these fountain pens and typewriters in their logos. I’m like, no, that’s all wrong. That’s not your trade. Putting letters on a screen is not what you do as a copywriter.

What you do as a copywriter is determining why are we doing this? What are we going to do to achieve these goals? Which inputs do we need? Like interviewing is like 90% of what you do as a copywriter. Bad interview, bad content, right? I mean, come on. And then at the other end of the production line, if you will, just check the finished product. Does this comply with what we’ve set out to do? Does it fit the tone of voice? Does it achieve our strategic goals? Does it resonate? All the marketing things you need to do and that typewriter, that needs to go. Just, I found, oh no, get it away from me.

Noz Urbina: Yeah, I think that’s really hard and I brought that from a whole different industry. I was, I brought that up. I got that one from my experience in technical communication and there’s a lot of people who were, you know, copy editing, for example, is a great example. I spell really well. I’m a, I can spot a typo that no one else spotted. And I’m like, that’s great. And I know that’s cool and you spent 30 years being proud of that and you should be proud of it. It’s like, it’s a cool skill, but there’s a difference between having a neat skill and having something that is marketable after a technical revolution because the skills mix that we need changes and nobody does just one thing that they do. You know, you have more value than that.

Bouke Vlierhuis: Yeah, absolutely. But the funny thing is once you start seeing these effects, you see them everywhere because like marketers, like marketing managers and managers in general, they’re the same because many managers will think that somewhere deep down, they will have the idea that what they do is write emails and update spreadsheets, right? That’s the inefficiencies of their trade are a lot of communication management.

Communication is just redundant. It’s people asking, what’s the deadline for this? So if you have your process in order, and there’s a board. Like, we have Kanban boards, we have them for decades, but B2B marketers never use them because they were making four white papers a year, so you would have a board with like four cards on it.

So when you start scaling up and when you start like making faster and smaller content, you need to look at that process as well. And you need to not be updating spreadsheets and sending emails, but that is also for a marketing manager and a leader or a manager in general. They also have the mission now to step back from their desk and look at what they’re doing and look at what is AI going to do for me, because copy pasting from Google Analytics to PowerPoint is something most marketing managers do every Monday that needs to go as well.

Noz Urbina: So full process rethink at every level of the organization. I think that, you know, your book also kind of provided some do’s and don’ts. So what are, we want to be aware of this, the chainsaw effect, avoid the fountain pen syndrome. Any other key pitfalls or tent poles that we can still rely on going forward?

Bouke Vlierhuis: Well, I think there’s five in the book, but I think the most two most important ones are, first of all, if it’s not working, don’t scale it up. So don’t use AI to scale up crap, because if you’re honest, like most B2B content is not working because it does not resonate and it does not resonate because it’s written and created from the inside out. So it’s about the product and not about the client.

So if you have four white papers a year now that are about the product and not about the client that are not being downloaded or not delivering any leads or prospects or sales or whatever, doing 12 or 24 of them or doing variations of them for every sector because that’s an easy thing to do with AI will not get you any results, so you will be scaling up crap. That is rule number one, just take the time that you win by using AI to actually think about what you’re doing, for whom you’re doing this and why, and to go look for these customer questions. So that’s rule number one.

Rule number two is human in the loop because AI content is not going to be as good as content that is written by a professional copywriter with 21 years of experience and not a lot of hair. So, and even if they do have hair, probably pretty good.

So there needs to be that final like 10% where you actually look at it and go, okay, we tried this hook last week and it didn’t really work. We need to go the extra mile. And then process thinking kicks back in because if you get a hook, like the first line of the scroll stopping line for a LinkedIn post that doesn’t work and you correct it, and then you get one tomorrow again, and you correct it again, then you have to go like, okay, I’ve corrected this twice. So that means I need to go back into the process and change what the agent or the workflow or whatever is doing because I’m doing repetitive work. So then you go back to process thinking. But the two most important things is don’t scale up crap and keep the human in the loop.

Noz Urbina: I’ve noticed that the not scaling up crap, you know, I was gonna stop you a little while back when you were talking about how you couldn’t do anything because of the creative creation bottleneck on the Omnichannel podcast. You know what we were, a lot of what we’re talking about is okay, yeah, it took you ages to create this stuff. Now, leverage, leverage, leverage. How do you reuse, how do you prevent creation from being a bottleneck by leveraging that asset you’ve created as much as possible.

And all what I’m seeing is that AI just pours gasoline over that whole thing because now you are both creating and leveraging, you have the capacity, let’s put it that way. You have the chainsaw now. So now you can really do all of this at an incredible speed. The human in the loop, it shifts things to strategy and quality. What are we mapping from user needs, questions over time, our understanding of the user, over to that user results? Are we getting people’s questions answered? Are they being solved?

And this machine that we’re setting up, and I use that term very metaphorically, the human and AI collaborative process that we’re setting up. How are we fine tuning it? I also was gonna kind of push back a little bit about throwing your prompts out every six months. I can, that’s another podcast, but I can talk about how to make sure that your prompts have longevity. But I think it has to do with making sure that you’ve thought through your process and that you have these chunks of independent logic, which makes sense.

So we know that breaking down an interview transcript should always look like this. So our rules for breaking down an interview transcript are consistent. Then there’s another question, which should be a different prompt. Like how do we take broken down nuggets of research and process them into insights? So at every step of your process flow and isolating it and treating it like an industrial process.

What I love about this AI revolution is it’s kind of forcing organizations to finally do what we should have, what we have been talking about for 25 plus years. Decades.

Bouke Vlierhuis: Decades. Yeah, I know. Yeah. And I love that about AI. Just like I said, AI has made process thinking and process analysis hip again, and I love it because that’s what I did for a living, like back then. And I actually have a client who has developed an ERP for selling, buying and selling used cars. Just because he was a car salesman for 20 years, and for 20 years he walked around, we need to do better. We need to do better. And I’ve been walking around in marketing for 20 years and thinking the exact same thing.

So I sent him the book and he took the trouble to contact me from his holiday address and he was like, “Yes, yes. Finally a kindred spirit.” So I think marketing is waking up. And part of it is because marketing, marketers are finding out they can automate things and they’re like, okay, we need to automate things. And they call someone who knows how to automate things. Like there’s all these people like doing no code and stuff. And the first thing these people ask is, what’s your process? And this is forcing them to think about, oh yeah, what’s my process? So I’m having fun. I’m having fun. You know, my whole business model was shot to hell, and I’m having fun.

Noz Urbina: Well, it’s interesting because I’m getting contacted by, and I used to talk a lot about the stretching of the market. That we would see more and more separation between the haves and the have-nots. And I think that’s, we’re finally seeing it play out.

So some people are racing ahead. You know, we’re having the same problem right now. Maybe you are as well. There’s, we just have more business than we can handle. Like everything’s going fantastic. And then other people are, there’s just this flood of job postings on LinkedIn, which is not always the fault of the people who are being let go. A lot of that time, that’s managers or organizations who assume that this is all just gonna work.

And as well, you know, we’ve been very positive about how this and how you can do it. Think about your process, but nobody really, we are in a period of transition, so we can make good predictions, but exactly where the chips are gonna land is still up for grabs. So, you know, a lot of people are getting let go on a hope or a vendor promise.

Bouke Vlierhuis: Oh yeah. Well, I, yeah, we all know about vendor promises. They’re just, you know, I think Gartner needs to attach an extra piece of paper to the top of their hype cycle for AI. It’s like through the roof. And I think looking at the things we’re reading now in the press and the things people are posting on LinkedIn, maybe it’s my bubble, but I think we’re heading for the trough of disillusionment quite quickly now.

Noz Urbina: So for those who are unfamiliar, Gartner publishes their hype cycle, which plots out the life cycle of hype for technologies. So you start, I don’t remember all the phases by name, but the big one, there’s the peak of expectations or something like that where the hype is like, it’s gonna be magic. It’s gonna change the world. We’re all gonna be geniuses. There’ll be no poverty. It’s just like John Lennon’s Imagine.

Bouke Vlierhuis: Solve climate change. That’s a big one as well. Yeah.

Noz Urbina: Everything’s gonna be fixed. And that’s where I would say that’s where we were six months ago.

Bouke Vlierhuis: I think we’re pretty much over the top with that. Yeah, absolutely.

Noz Urbina: We are heading down to what Gartner calls the trough of disillusionment where people recognize that all of the promises that they heard are not coming true. And then it gets very negative. What comes then is, I don’t remember what they call it, but we slope…

Bouke Vlierhuis: The plateau of productivity.

Noz Urbina: The plateau. Oh, there’s another one before that.

Bouke Vlierhuis: Yeah.

Noz Urbina: You know, we have to then ramp back up from, oh, that was a waste. You know, bunch of lies. The vendors were all, it was just a bunch of junk, to realize, oh, actually there’s a lot of value here. But what are the processes? What are the systems? What are the best practices? What are the expertise mix? What does the reorganization of teams look like? And when we do that, then we ramp up into eventually to the plateau of productivity, which I think we’re far from with AI. So how do we people, how do we people get on that ramp?

Bouke Vlierhuis: Well, the thing I’ve been doing for the last couple of weeks, month maybe, is I’ve been just, you should try this, talking to people, which is pretty cool. I didn’t know this existed, but I just sit down and I talk to an actual person about what they’re actually doing and not doing with AI and if they’re not doing things that I think they might be doing, ask them why they’re not doing that.

So what I’m collecting by doing that is first of all actual practical use cases because there are actual practical use cases. I’m the prime example. You know, I’ve completely transformed my business model. I am an interim marketing manager now for two different companies at the same time, while also doing their content creation. So AI is helping me work two jobs in the same hours. Well, not to brag, but you know, that reflects on your income. Let’s just, you know, so I’m a successful AI use case myself, and I talk to people.

I talk to a software engineer who has fully automated compliance documentation and solution design for cloud applications like fully automated. Like five person team, spent two weeks kind of work. And I’ve also started, and this is also fun, talking to people who don’t work in marketing at all. People who work in HR, middle management, middle management has like the most awesome AI use cases like that goes from not watching a 30 minutes video. The senior management sent you because they have read somewhere that you should update your people by talking to a camera for 30 minutes. I don’t know, not watching that, but summarizing that to writing a speech for a retirement party for someone. Because you know, managers are like really fucking busy, right?

So that’s expected of them, and they want to do it right. They want to be empathic and all that stuff, but they’re busy. They’re really, really, really busy. And I’ve been compiling like, well, I don’t want to call it a database because it’s just like a bunch of notes, but of actual successful profitable AI use cases. And there is a lot of them out there.

Noz Urbina: So is that your next white paper? Is that your next book? I want to see the use cases. Show me the use cases.

Bouke Vlierhuis: Yeah, well, I want to write them down, but I want to have at least 30 conversations with people, so I’m not even halfway yet, so that’s gonna take a couple of months, but I want to write this down and when we hit that low, where everybody’s going like, “Yeah, yeah, I told you this AI stuff’s not gonna work. Just, let’s start typing again.” I want to show people that this is actually working. It’s just not working as quickly as the bros were selling it to us.

Noz Urbina: I think it’s been sold these promises and I, we’ve been watching for 20 years. Things just happen faster and faster. So I remember, I always think of those adoption things. So like how long did it take for everybody to have a TV? How long did it take for everyone to have a VCR? How long did it take for everyone to have a CD player, DVD player and the curve just keeps getting sharper and sharper and then it switched to apps, so you didn’t even have to get a physical product.

So how long before everyone was on Facebook? How long before everyone had Instagram, blah, blah, and then TikTok and then ChatGPT was the fastest we’ve ever seen since that users in record time. And so we kind of get this idea that everything should just happen like that now. And if it doesn’t just happen, if it takes two years, then that’s glacially slow and obviously a failure.

Bouke Vlierhuis: Yeah, but you have to remember that we’re talking about enterprise, like corporate adoption here. So what, now we’re not human beings adopting in their life and the small businesses or their enterprises, if they’re humans facing real change. Not just, oh, this is just a click into my regular life, but real change that takes time. And especially, well, yeah, and we’re still getting used to like working from home and having Teams meetings. Like there’s still no emotional social framework for working from home and talking to each other in these little digital squares.

I keep asking everyone, what do you do if someone cries on a Teams call? Right? No one knows. No one knows. So we’re emotionally adapting to like the last two or three technological revolutions and this AI thing, you know, the technology is there, like the use cases for GPT-4 and for GPT-5, the difference is marginal. So I think technology wise, we can do anything we want right now.

So adoption is like the next frontier. So if you’re 20 years old and you’re listening to this and you’re wondering what career you should pursue, adoption consultants, this is going to be the gold mine. The next like two decades, help people actually use this and deal with the change that it requires of you.

Because the change is huge and I’m 49 years old. You know, I miss typing for a week and delivering a white paper. I miss it. I feel these emotions and I’m like half tech bro and half like 49-year-old copywriter. So people always think that these things are easy for me because I move pretty quickly, like business-wise, but this thing that happened to us is huge. So don’t, you know, if you’re a manager and you think your people are slow, you know, don’t scream at them because they’re adapting. This thing is really, really, really big.

Noz Urbina: Help them, facilitate them.

Bouke Vlierhuis: Help them. Yeah. And that’s what you’re, you know, did I show you this one? I wrote this to help B2B marketers, because B2B marketers were overwhelmed, like pre AI because there’s, you know, I work for a company that has 50 salespeople, 50 salespeople, and two marketers, and this is all through B2B. This is their daily fight. You know, they want to service all this entire, you know, the sales machine, understaffed, underfunded, underskilled usually, and then AI happens. So this book was written to help them.

Noz Urbina: A survival guide. And I can’t agree more about what you said about adoption consultancy. So we, you know, I’ve been doing trainings and workshops for my whole career, our AI adoption workshop is nuts. Like, I’ve not seen people react to anything like that in my career.

People are happy and we’re, you know, it’s so in demand. And I also, not to brag, but I got like that email, that email that every workshop leader wants to get after our workshop where someone says and said, we just went from an already understaffed team of 20 down to three. Fired everybody, leadership decided to go all in on AI. And then it, like not in all caps, but in bold type in the middle, I truly believe taking your workshop is what helped me stay employed.

Like, you know, either the people who can adopt, the people who kind of, rather than just kind of fight it and point out all the flaws. Point out the flaws, but go, okay, how do we work with the flaws? Because it is flawed, which something, you know, we haven’t touched much on this podcast, but you can check out our many, many other episodes. We, nobody is saying that AI is perfect. Nobody is saying that the, none of the flaws you’ve heard about exist. They absolutely do exist. The question is how do you put in process so that you can work around those flaws. And you mentioned human in the loop earlier, how do you build a human AI process that is robust if you build your process so that it has zero tolerance for flaws? Bad process.

Bouke Vlierhuis: Absolutely. Absolutely. And, you know, AI just sometimes it’s just does these weird things. So people say, AI, you should treat AI like an intern. Well, I think that’s an insult to the interns, basically, because you know, sometimes, you know, a 700 word blog post, that should be like an easy machine readable instruction right? Well, that’s…

Noz Urbina: That is a funny one. So there, I’ve worked with a bunch of interns and I do get stuff from interns and go. What, what, what? Just, you know, you forget, you know, 25, 30 years into your career what it was like to be 22 and not have the context. It’s not like you’ve known nothing, but it’s tempting to say that. It’s tempting to say, do you know nothing? Like, because you’re just new. You’re just like, oh, I just didn’t know that. I didn’t have that background, didn’t have that context.

So they sometimes they come with stuff where, yeah, it’s, for me, it’s a very similar experience to what the AI does. But you’re right, there is this little area of failure modes, which are just not human failure modes, like no human… I’ll give you a perfect example. We were working on website content and the prompt said, you know, write up, draft up this piece of content and add three related links to case studies from the website. And if it didn’t have case studies that were on the website, it just made them up.

Bouke Vlierhuis: Yeah. Yeah. I have like white paper templates that say insert reference cases here, so you have to delete that from your template because it will create the reference cases.

Noz Urbina: And then I had this person searching the website going, where are these case studies? I can’t find them. And I’m like, because they don’t exist. Yeah. Yeah. We’ve never worked with Procter & Gamble. They just never happened.

Bouke Vlierhuis: No. Oh, ask me about this in-depth research about AI in the artificial, the audio visual industry that never exists.

Noz Urbina: Actually we should talk about that, the MIT headline, that’s all the buzz. 95% of AI POCs or projects fail.

Bouke Vlierhuis: Yeah. Yeah, I, yeah, sure. Yeah. Innovation fails, right? And most of it fails because the expectations are too high. And the other things that fail should fail. You know, well, you win, you win and you lose, you learn. That should be like the mode we’re in. I think especially with agentic AI, I mean, the things that can go wrong in an agent project are just, there are failure modes there that we don’t even know about.

And we should, you know, probably hurry up and test things and find these failure modes and understand them. Think of ways to work around them because I think that most of the things we’re doing with a technology we knew nothing about, like two years ago, are failing. It…

Noz Urbina: I think that they shouldn’t fail. I think there’s some context to that 95% number. That’s 95% of the time when the management just pushed out AI onto everybody and said, go be productive. And so, yeah, here’s a tool. You have no training, you have no background. Be productive. And then they failed to meet expectations. No, duh.

Bouke Vlierhuis: Yeah, I’m an aerospace engineer. You know how we achieved like the safety level of airplanes that we have now? Because they fall out of the fucking sky all the time. And then we go like, hmm, maybe, yeah. Well, there’s, yeah. This is how we discovered like metal fatigue. Because aluminum airplanes were falling out of the sky. So yes, obviously the first couple of years things are going to crash and burn. Of course, of course. And you should budget for it. And you should build an infrastructure.

And there’s actually a whole chapter in the book about thinking about innovation, how do you think, how you get your head around innovating and failing and building an infrastructure that takes failure and weaponizes if you will, the lessons from it and injects them back into the innovation process. Because you’re not just in your process process, you’re always also in an innovation process, which is innovating your process process.

So you have to, as a marketer that used to like send out briefings on Monday and wait for the white papers to come in. And now you’re just a process person, which also involves innovating your process and accelerating it and integrating new technology because new technology lands about every minute and a half right now. So…

Noz Urbina: Love that. Okay. Well, we are sadly at time. Thank you, Bouke, for joining us again and contributing this book to the community. I hope if you can read it in Dutch, please do check it out. Otherwise, maybe a translations sometime in the future to watch this space. Otherwise, thank you all for listening and we will catch you on the next one.