Modelling Content for EdTech: 27zero’s Customer Stories Lead Sebastián Pulido on Writing Strategically (with AI)

Mr. Pulido, thank you for being part of EdTech Mentor. This is going to be a fascinating conversation because, during the preparation for this interview, we discussed how copywriting and writing in general are the foundation of many channels and initiatives. Writing essentially brings campaigns and projects to life, making it a crucial part of what we do at 27zero.

Let’s start with your journey—how did you get here? Tell us a bit about your education and background.

Thank you so much! It's truly an honor. I've worked in EdTech for almost my entire career. I studied Communications at the Universidad Javeriana in Bogotá, with a focus on editorial work, and originally thought I'd go into more traditional media. But during my studies I became more and more interested in digital media.

I remember watching things change almost daily—every month, every semester, everything felt outdated. The most valuable knowledge I gained came toward the end of my studies when I started doing more hands-on work. Today, that’s even more true. The industry evolves at lightning speed.

My interest in EdTech and digital media grew towards the end of my studies. I landed my first role in EdTech during my internship, working for eLearn Magazine, an initiative of Blackboard. That's when I fully immersed myself in the field. I didn't meet Laureano at that time, but I know that he was already leading the project through Nivel Siete.

Right! The original name was The eLearner, but when we went to register the brand, we realized it was already taken somewhere in the world—something we hadn’t foreseen. So when we decided to launch it globally under Blackboard, we had to rename it eLearn.

So, your background is rooted in journalism, and this reflects a broader trend. In Colombia, many communication students naturally gravitate toward journalism, though not all, of course. Would you say you followed that path?

Not exactly in the pure journalism sense—some students do focus on that. But my emphasis was more on print media and editorial work. I originally thought I’d dedicate my career to that. And in some ways, I still am, just in a different form. I’ve never really left the editorial world—it has just evolved.

That’s a perfect segue. I’d love to hear your authentic take on your first deep dive into the EdTech world through eLearn Magazine. When you started writing content and working on that publication, what was your first big reality check? And what year are we talking about?

Not that long ago—this was around 2017.

Right, the magazine was in its prime then.

Back then, I didn’t even know what EdTech meant. Of course, I understood "technology in education," but I had never thought about it as a defined industry. I had always been drawn to technology, but more from the entertainment side—video games, media, and so on—not education.

Entering this space felt a bit intimidating. Up to that point, my approach to technology had been organic—I followed what interested me, what seemed innovative. But suddenly, I was immersed in advanced concepts related to communication, technology’s impact on interaction, and even how it affects us physically and socially. These topics always fascinated me, but never specifically in education.

Then, I had to start learning about pedagogy. It was a highly academic environment—often, the Ed part of EdTech dominated decision-making more than the Tech. In marketing, the tech side is ever-present, but it operates differently than in any other tech sector. You have to learn an entirely different framework and mindset.

Makes sense. Do you remember your first assignment for eLearn Magazine?

That’s tough—I had a few at the start. But my first major piece was about Blackboard Learn, back when Ultra didn’t exist yet.

I was interviewing people from all over—one person from the Netherlands, another from the UAE, and several in Latin America. I don’t remember a specific one, but I clearly recall that initial phase. Everything was so new, and in every interview, I discovered something different. I had to keep learning and figuring out how to speak about these topics with enough depth to create valuable content.

For context, eLearn Magazine was originally run by Nivel Siete, then acquired by Blackboard, which later became part of Anthology. Now, eLearn Magazine operates under OpenLMS. Anyone interested in the full backstory can look it up.

But what’s important is that, at its peak, we published around 450 pieces worldwide with voices from this industry. And back then, people didn’t even call it EdTech—it was simply "technology in education." It was a massive operation with some incredibly interesting people.

Exactly. The concept of EdTech was still forming back then. Today, it’s a recognized keyword, but at the time, it was more informal—used internally but not as a standard industry term.

That brings me to a related topic. Even though we specialize in this industry, every time we onboard a new client, we face the challenge of understanding their audience—their buyer personas.

You manage multiple clients at once while keeping track of their industries, products, and audiences. How do you keep all that straight? And how do you ensure efficiency in tailoring messaging for each one?

You can’t rely solely on memory for that. Your brain should be used for other things. Instead, I depend on systems and structured processes.

During onboarding, if done correctly, we dive into these aspects with the client—who their buyer personas are, what their core offerings are, and whether they should refine their focus. Sometimes, they come in wanting to promote an entire portfolio, but it turns out that emphasizing one or two key products is more effective.

These personas can even shift mid-campaign. We may discover alongside the client that certain things aren’t resonating, leading to adjustments on the fly.

Exactly. That’s the messaging and positioning evolution. You might define a positioning strategy early on, but the messaging—the way it’s framed and refined—evolves significantly as you interact with the market.

Right. In EdTech, sometimes the process is straightforward, especially with established tools like LMSs or SISs. In these cases, we know the decision-makers we need to target.

But even within similar institutions, decision-making varies. Some universities empower academic leadership in EdTech purchasing, while others see the university president taking a more hands-on role. In some cases, the buyer persona is a single person with a name and a title. That dynamic shifts depending on the institution’s size, type (public vs. private), and procurement process. It’s not one-size-fits-all.

And yet, you manage to keep track of all of them, which I truly admire. On that note, without naming names, can you share the most challenging client we’ve had? Maybe just mention their industry or product category. From a messaging and positioning standpoint, which one has been the toughest to handle so far?

The most difficult type of client is always the one that comes to the agency because they aren’t entirely sure what they want to say or sell—but they don’t realize it.

Of course, as an agency, we understand that clients sometimes have uncertainties and come to us with those expectations. But at the end of the day, it's a different kind of service. It's closer to product marketing or even marketing consulting.

Yes, that’s exactly product marketing. Once that’s clear, everything makes more sense. Good answer.

And what about the easiest clients, Sebas?

Ironically, the easiest clients might also be unsure about what they want, but they’re open to collaboration when something isn’t working.

It really comes down to attitude—understanding what an agency is for and why they sought one out in the first place. Some clients don’t have a clear direction but still see the value in working with an agency and leverage that collaboration.

That makes the process much more open-ended and cooperative. We know EdTech, but as I mentioned earlier, it’s an industry that evolves incredibly fast. Expertise in this space tends to be siloed—people might be extremely knowledgeable in one area but less so in an adjacent field, and that can create friction.

For an agency, that’s a challenge. Even though we’re highly specialized in EdTech, there are many different types of solutions touching vastly different aspects of the industry—scholarships, financial aid, purely academic tools, and so on.

So even within a niche focus, there are broad subcategories to navigate. That’s why working closely with clients is crucial. They’re the true experts in their domain, and our job is to extract the insights we need to help them communicate their message effectively.

Got it. At the start of this conversation, I mentioned how messaging and copywriting touch nearly every channel we work with.

It’s not always obvious to people reading this, but content marketing involves different types of assets. There’s content engineering (which we’ll get into in a moment), social media, email marketing, ABM campaigns—and within all that, advertising plays a huge role.

Words are central to everything. From social media copy to events, messaging is key. It all starts from an overarching positioning and messaging strategy, but then it gets broken down into different formats.

So, what’s your favorite format to work on? The one you truly enjoy the most?

I don’t really have a single favorite, but I’d say I enjoy working on whatever format best fits the purpose and audience.

The effectiveness of a message depends entirely on the campaign goals and the buyer persona.

That said, I love writing video scripts. I also enjoy working on ads—figuring out how to build compelling ad copy around keywords is fascinating to me.

Formats can range from short, snappy ads to in-depth interviews or case studies. A campaign might combine all of these—an article supported by video and ads promoting the content.

What I enjoy the most is when a piece actually works—when I can tell that it’s serving its intended purpose and resonating with the right audience. That’s what makes my brain light up.

Fair enough. Now, I’m guessing you’ll say no, but is there a format you struggle with? One you find the most frustrating?

At first, social media copy. I understand it better now, but in the beginning, it felt repetitive if you didn’t approach it correctly.

It can become background noise.

Exactly. You have to be really intentional about it.

Just because social copy is short doesn’t mean it’s less work. In fact, the less text you have, the harder it is—because every word carries more weight.

Crafting the perfect LinkedIn post, for example, is crucial. The audience is high-level, with highly qualified leads reading it. There’s a whole brand voice and positioning strategy behind it, built up over time.

Anecdotally, is there a project or a piece of content that stands out as your favorite? The one you remember most fondly?

It has to be Legacy Conversations. It was my first major project with 27zero, and it was such a rewarding experience. It was complex, intense—I put in a lot of hours on that one.

We both did! I conducted most of the interviews myself. I’d say at least half of them.

Now that you mention it, those interviews covered both professional and personal aspects in a very natural way. That’s why they resonated so much—not just for the interviewees but also for the readers.

Both the interview process and the final product—the illustrations, the writing—were incredibly detailed and well-crafted. That level of depth is rare to see today.

We were interviewing university presidents—people who are often placed on a pedestal. Academia has this traditional aura of prestige and hierarchy.

But we showed a more personal side of these high-profile figures. I don’t think that had really been done before in the EdTech space at that level.

The impact was huge, not just for the presidents themselves but also for the broader EdTech community. The reception was overwhelmingly positive.

And it felt warm, which is rare in B2B marketing.

Moving on, we’ve worked on major projects where we’ve had to bring in additional talent—copywriters, translators, and other specialists. Even outside of writing, we’ve collaborated with designers, video editors, and more.

You’ve been responsible for hiring many of these people. How do you recognize the right talent? How do you know they have what it takes to deliver on a project?

Like in any hiring process, there are some basic criteria—credentials, experience, a strong command of English. Those are the fundamentals.

But beyond that, the key is to test the waters.

If I believe someone has potential, I’ll assign them a small, non-critical piece to see how they perform. It doesn’t have to be a full test or a rigorous exam—just something manageable that fits into a real campaign.

The tricky part isn’t the evaluation—it’s finding people with the right mix of skills in the first place.

Hiring for EdTech marketing is particularly challenging. There aren’t many people who have done this before, at least not in a specialized way. But if someone is eager to learn, they can develop the skills. It’s not an impossible field—it just requires the right mindset and effort.

Speaking of scale—especially when it comes to larger projects and content marketing—before diving into content engineering, which I really want to cover, let’s talk about the generative AI revolution.

We experienced it head-on because this technology became incredibly powerful, incredibly fast. Of course, some people mistakenly believe they no longer need writers. But in the effective campaigns we run today, how do you use these tools?

I know there are several, but give us a quick glimpse into how AI helps you achieve results for clients while still relying on strong human input.

At this point, almost everyone at the agency is using some form of AI—we’ve been doing it for nearly three years now, ever since these tools became widely accessible.

As the tools have evolved and improved at such a rapid pace, our perception of them has also changed. So has the way we use them and the role they play in our workflow.

Looking back, when we first started using AI, we had the same expectations (and fears) that many people did—that it could do everything perfectly and that we might be replaced.

AI can do a lot, and it can be incredibly useful, but we quickly realized that the quality of the output is directly proportional to the quality of the input. The knowledge and judgment you bring to crafting a prompt determine what you get back.

That learning process—figuring out how these tools actually work—is what’s brought us to where we are today: AI is an indispensable tool, but it’s not omnipotent.

Take transcription, for example. We use it a lot, and it has improved dramatically.

When we first started, I remember early transcription tools required so much manual correction. They were helpful, but the cleanup work was significant.

Now, AI-powered transcription tools are so accurate that a final review doesn’t require nearly as much effort.

These kinds of automated, repetitive tasks—things that used to be time-consuming—are where AI really shines.

That said, we also have to understand its limits. A standard transcription now takes as long as the tool needs to process it—because more than 90% of the time, the output is spot-on. It even picks up proper names, industry terms, and event names, which it then cross-checks.

It still amazes me how much it has improved. Some tools even translate transcriptions automatically.

But when it comes to content generation, you have to know what AI is and isn’t good for—because it’s easy to fall for AI’s “hallucinations.”

That’s exactly what I was about to say. You’ve made several key points already, but one of the most important things to understand is that AI models are specialized intelligences, not general intelligences.

Exactly—especially when it comes to research.

I remember when we first started using AI, we wanted to create some more technical white papers. I thought, “Let’s see how well it can find sources.”

Of course, that was a disaster. Large language models don’t actually retrieve or verify sources. They’re just predicting the most probable next word in a sequence.

If you ask for a list of references, you’ll often get completely fabricated citations. That’s where AI’s “hallucinations” come in.

And the problem is that AI delivers these fabrications with total confidence. It just asserts them—no hesitation, no qualifiers.

The references may sound plausible. It might list real authors and real topics—but the actual publications don’t exist.

Honestly, that reminds me of some business leaders I’ve met. The resemblance is uncanny. (Laughs)

And in B2B marketing—where credibility is everything—that kind of confident misinformation is a nightmare.

Alright, let’s shift gears and talk about content engineering—how we work with it, why it helps, and your experience with it.

When does it really make sense to scale content using this approach, and when doesn’t it?

I think our conversations about content engineering really started when we saw the potential of AI—not necessarily because everything should be AI-generated, but because we wanted to optimize processes.

Our goal is to make content production not only more efficient but also more versatile and useful—both for the agency and for our clients.

Because once the expectation emerges that AI can do everything, the expectation for content volume also skyrockets.

So, AI isn’t taking work away from us—it’s actually creating more work.

Exactly! AI is giving us more work.

Just to give a little context before you continue—content engineering is all about figuring out how to produce content at scale.

It forces us to think about composability—how content can be structured in reusable, modular ways.

And, as you mentioned earlier, generative AI is only as good as the information you feed it. Content engineering helps ensure that any AI-assisted content generation is grounded, avoiding hallucinations, misinformation, or inaccuracies.

Right. By leveraging these tools, we started building out a content engineering approach.

Essentially, we map out all the possible content a client might need—across campaigns, across formats, across different use cases.

That includes everything from website copy, product descriptions, marketing materials, sales collateral—you name it.

We analyze what information each piece of content depends on and break it down into its smallest meaningful units.

Once we’ve reduced content to its core building blocks, we can then reassemble those blocks to create new content dynamically.

This allows us to generate a variety of assets—from long-form product descriptions to short, snappy CTAs—while maintaining consistency and efficiency.

Actually, let me pause for a second.

I think it would be really useful to include a diagram of a typical content model here. We can anonymize it, of course, but adding a visual representation would help illustrate these content building blocks and their relationships.

That way, readers can see how structuring content this way makes it composable, reusable, and scalable.

It also shows how this approach allows multiple people—not just one or two content creators—to manage a high volume of content efficiently.

And that’s another key takeaway—there aren’t many tools that let you do this kind of end-to-end content modeling easily.

From building the content structure to actually generating the final content assets, there’s a gap in available solutions.

Most existing tools are designed with web content in mind. But content today is deployed across so many different channels.

The best platforms we’ve seen—like Contentful—are still very web-centric.

That’s where the agency’s role comes in.

We manage the entire content engineering process and then distribute content to creative teams, ensuring every piece is aligned with the broader strategy.

Exactly. Alright, cool. So, when do you think it does make sense to dive into content engineering, and when doesn’t it?

That’s a great question. Based on our experience, content engineering makes the most sense when a client has a big vision for what they want to do and understands the long-term value of content.

If they’re truly committed to marketing their brand and products over time, it’s a great investment—because once it’s set up, it’ll serve them for as long as they need it.

But if it’s just a one-off campaign—like driving interest in a specific event—then it’s not worth it.

For short-term efforts, the content is usually very straightforward: a booth setup, some marketing materials, a few emails, flyers—traditional stuff.

But if a brand wants to establish itself and has the vision and commitment to grow long-term, then building a content model is absolutely worth it.

Got it. I want to ask about the cultural aspect of our accounts.

It’s no secret that 27zero was born in—and is largely made up of—a Latin American team. But today, almost all of our clients are from other parts of the world, mainly North America.

You have an incredible number of cultural references from around the world. My question is this:

Of course, you use that knowledge in your day-to-day work. But why do you enjoy it? Not everyone has that kind of curiosity.

What drives you to explore so many different cultural perspectives? What’s the root of it?

It’s a deeply personal interest.

For me, it comes from a need to understand things—understand people. Why do they think the way they do?

It’s almost paradoxical in a way. You start by looking at how different someone is from you, but to truly understand them, you can’t focus on the differences—you have to look for common ground.

That’s what makes it easier to connect with people.

When you understand cultural differences—and the unique perspectives people bring—you approach those differences with more empathy and awareness.

That’s a great answer.

Here’s another important topic we haven’t touched on: client feedback.

With everything going on today—the short attention spans, the constant shifts in messaging—it’s easy for even well-established brands to lose their way.

Even within the agency, keeping that strategic focus isn’t always easy.

You get a ton of feedback all the time.

How do you take a step back, breathe, and integrate feedback intelligently?

How do you filter what’s valuable and keep everything aligned?

I try to reduce feedback to its simplest form.

If it’s a factual error, that’s easy—it gets fixed.

But if it’s messaging-related, I go back to the basics.

For instance, I refer to the content model we’ve built. If a message was already approved but is now getting pushback, that’s a sign we need to reevaluate:

“Okay, maybe this message isn’t as strong as we thought.”

That’s one of the biggest advantages of content models—when you adjust something at the foundational level, it automatically updates across all related content.

So if a client feels a message isn’t landing right, we can step back and ask:

• Why doesn’t it work?
• Is this just a one-off revision, or is it part of a bigger shift in messaging?
• Are we changing the way we communicate as a brand?

If the change is minor, no problem—we adjust and move on.

But if it’s a strategic shift, it’s worth a deeper discussion.

And that’s why aligning on purpose from the very beginning is so important.

If we all have a shared understanding of what a piece of content is supposed to accomplish, there’s a lot less back-and-forth later.

But if we’re seeing frequent messaging changes, that’s a signal to ask:

• What’s driving this shift?
• Is this an isolated revision, or does it reflect a broader change in strategy?

It all depends on the relationship we’ve built with the client. If we have a strong working relationship, we can ask:

“Why are we making this change? Does it align with how we’ve been positioning the brand?”

If not, we need to reassess before making major shifts.

And that’s expected, right? Like we said at the start of this conversation, messaging naturally evolves over time.

This was excellent, Sebas. Thanks so much.

No, thank you. This was great.