Usama Khan (00:00): Thanks so much, Kate, for joining today on the VISIBLE podcast. Before we get into it, tell us a bit about yourself. What do you do, and how did you end up building a content agency specifically for B2B tech?
Kateryna (00:11): Sure. My name is Kateryna, and I run Zmist & Copy. It's a content marketing agency focused on creating quality content. We take care of positioning, strategy, and content production. I started back in 2013 when I got hired as a writer at a software development company. I stayed there for about four years, jumping from writer to chief marketing officer. When I started, they were three people. When I left, they were just over a hundred. Then I started another agency, more SEO-focused, called CAPS, with a co-founder. After four or five years we parted ways, and I opened a new agency, Zmist & Copy, which I currently run. It's in its third year now.
Usama Khan (01:18): Amazing journey. You mentioned you started an SEO agency. With Zmist & Copy, are you currently only doing content, or also SEO strategy?
Kateryna (01:28): We do SEO strategy, but we're not SEO experts and don't focus specifically on SEO. We position ourselves a little differently from most SEO firms, because we don't use some of the common approaches in the tech industry. We know how to create an SEO strategy, how to track where you are, how to check your visibility and what influences it. But we're not specifically an SEO agency. We often collaborate with SEO experts on the client side, or with SEO agencies directly.
Usama Khan (02:07): That brings me to my next question. B2B tech companies often treat positioning and search visibility as two separate workstreams. One team figures out the messaging, another writes the blog posts, case studies, and product pages. You start with positioning before any content gets written. Why is that, and how does that foundation change what the content can actually do for a brand?
Kateryna (02:27): It comes from my experience working with companies. When things don't work out, it usually traces back to problems in their positioning. Especially now with AI search. To be visible when somebody asks a question on ChatGPT, Gemini, or Claude, you need to be very specific about who you target and what you do. You need to be narrow.
The problem I see in almost 90% of cases is that companies are very broad. They have a vague understanding of their target audience. They say, "We work with a bunch of different companies, and we can provide value to all of them." That can even be true. But you cannot win fast in marketing without a lot of investment if you take that broad strategy.
That's why when we start with clients, we first figure out what makes them different and what exact ICP they have. ICP is a narrow thing by definition. You can have ten ICPs, of course, but to win at first, you need to position yourself in a very specific market, build a content strategy around that, and get results. If it works and the audience seems too narrow, then you can broaden a little. That's how it started. Right now, most of our engagements start with positioning.
Usama Khan (04:21): You made a good point about nailing positioning and writing specific content. What I've noticed is that when a buyer asks an LLM something like "what's the best project management tool for a 20-person remote engineering team," the model picks from hundreds of options. The brands that get recommended tend to have the clearest positioning around that exact use case, plus a branding element working in their favor, like more brand mentions across third-party sources. The bigger the brand, the more benefit you have anyway. But do you think positioning clarity is becoming a direct input into whether LLMs recommend your clients?
Kateryna (05:04): Yes. It's important to understand positioning first for yourself and your team, so everybody tells the same story when they meet a client or describe what they do. Then you document that positioning on your homepage and build your website structure from there. Everything on your website goes back to that positioning. When that structure is well thought out, and all your articles, content, and LinkedIn posts tell the same story, it really works.
You made the example with project management tools. Right now we have a lead, not a client yet, with exactly that problem. The company has a project management tool, and project management is such a broad field. I asked how many competitors they have, and they said more than 300. So how are you going to win, when there are very big, dominant players in that market? In that case, you need to position yourself very narrowly. It's the only way.
Usama Khan (06:15): That's an issue that usually comes up with broad categories like project management or field service. My next question: your content starts with expert interviews, and I believe the quality of content depends on the quality of those interviews. What's your process for designing questions that pull out something truly unique, the kind of insight, framework, or data point that only your client could share? And how do you recognize when you've hit one of those moments versus when you're getting generic answers?
Kateryna (06:49): Great question. Whatever we do in content marketing, we always start with interviews. Even when we figure out positioning, that's a two-month exercise with weekly calls with the whole team. All the company stakeholders are present, not just the founder and CEO, but also the marketing manager, head of sales, head of delivery, and head of technology.
By the time you're through that exercise, you know them pretty well. It makes the following steps much easier, because you already know what the company does differently. Sometimes they need help branding it. They have a specific approach and can see it's different, because we analyze competitors when we do positioning, and we have a lot of experience in the tech field, so we know how it works. We can tell when a company does something very differently. Our job is to wrap it into something that works in content. That means creating a framework, some signature IP, a name for that approach that's unique and differentiates the company.
When you do the interviews, it's important to come prepared. You need to know what final result you want. But you also need to leave room for free expression. It should feel like a conversation, not an interrogation or a Q&A session.
This is the mistake many writers make. They prepare a list of questions. I've been on calls with other writers who shared their screen and went through every question one by one. The person stays concentrated on the questions and can't go anywhere else, can't tell you an unexpected story. That's wrong, because you still can't be in their shoes. You need to leave room for unexpected things.
And you need to ask for a lot of examples. Whenever you ask something like, "When you have this problem to solve, what's the common way, what are the common wrong assumptions clients have about it," they need to give you specific examples. The specific examples are what make your content unique. You can also ask them to tell specific stories. "Tell me about something that happened this month, or last month, something you remember."
It's also good to ask questions that are a little negative, or rather, something they don't like. "What do you hate about clients looking at that problem? What do you always push back on?" When you get answers to those questions, that's what makes content interesting and even emotional.
Usama Khan (10:13): This was so insightful, Kate. What I've also seen is that story-driven narratives make content insightful and memorable, because people resonate with people and stories, the problems they faced and how they solved them. Another takeaway was the framework-naming point. It instantly makes content more shareable. If you have a process and you name it, it becomes more memorable, and it becomes easy to refer to when talking with a colleague.
Kateryna (10:50): Yes, exactly.
Usama Khan (10:51): Something I've seen on the AI search side is that content built from sales call analysis, from how your ICP actually describes their problem, shows up more in LLM responses. Real buyers don't type "what is the best field service management software." They type something like "we're a 30-person team currently using spreadsheets, we want field service management software that integrates with QuickBooks and has built-in route optimization." They give more context based on their situation, their tool stack, and their constraints. That specificity only comes from sales call analysis. Have you seen this play out with the content you produce?
Kateryna (11:38): Yes, we've been saying that a lot. What I'd add is that for AI search, everybody wants to be cited. To be cited, you need something unique in your content. The best way to create that uniqueness is to build a report, because it requires real effort. You go to people, talk to them, gather their replies. That's very unique.
If you're not ready to invest in reports, you can create a regular blog post, even one about how to do something. But in it, you describe your approach, your stories, your specific approaches, your named frameworks. All that signature IP is what makes you citable. When somebody asks a specific question, the model can cite those things if they're described well. It can say, "This company has a thing called this framework, and here's what it looks like." That's what I mean by named frameworks. We're all trying to create uniqueness. It's no longer enough to run an anti-plagiarism check to see how much you've copied. You need something actually unique there.
Usama Khan (13:09): That makes sense. You mentioned being cited. I've always been vocal about the difference between being cited and being recommended, the LLM actually recommending your brand as the solution to the query the searcher types in. Is that a differentiation you make with clients?
Kateryna (13:29): Yes, of course. It all contributes and compounds over time. You want to be mentioned for consideration-stage queries, like "what's the best company for that." To be citable there, you need to be specific about your positioning. You need that positioning nailed on your homepage, blog posts that speak about the same thing, and landing pages describing all of it. That's how you get citable. You also get cited when people ask awareness-stage questions, when they're just researching. It all compounds. I believe that's the right approach to content today.
Usama Khan (14:19): When the goal is to write highly differentiated content that stands out in AI search, where do you think AI is genuinely useful in the production process, and where do you draw the line?
Kateryna (14:32): Another great question. AI is great at helping you build briefs, but only if you've fed it enough context. Previously everybody talked about prompt engineering. A year ago it was even a thing. Now it's no longer about prompt engineering, it's about context engineering.
What is engineered context? It's all the positioning, the target audience, the ICPs, the insights from sales calls, and your strategy. If you have a well-documented strategy that describes who you are, what you do, what you want to achieve, and what messages matter to you, and you feed it into AI, then whenever you need a brief or an outline, it can help you well. It can also help with research. Claude Code is pretty good at gathering information online, crawling and so on. If you need to pull insights from forums where people share opinions, it can organize them in a table, and you can draw insights from that. That's great too. So research, briefs, and some rewriting or rephrasing when you're stuck on how to express something. That's my top use case.
Where AI cannot substitute you is the real stories, the real experience, the things that make you you and not a robot. A unique angle you want to describe. You can have a piece written in an AI voice, but if you have an interesting insight, it can still work, because that idea comes from your thought, not from AI.
I hate seeing all these AI-generated posts. I get triggered, because I know the specific phrases and constructions AI uses. But if there's an interesting idea there, it can still work. The idea has to come from you. You don't just say "write a brief about this." You say "here's what I think is most important to focus on in that brief." That's what can't be substituted by AI.
A funny example. My sister tried to write an article a couple of weeks ago using ChatGPT. She gave it to me to read. The idea was interesting because she knew what she wanted to say, but the way it was written was very triggering, very clearly AI. I told her she needed a lot of edits. She produced another draft and told me she'd run it through the Grammarly humanizer. I reread it, and that draft looked even worse than the first one. It was maybe less triggering in terms of AI, but it was way weaker. A human being can edit what AI produces. AI alone can't get there.
Usama Khan (18:11): Very well put. I personally love Claude. I always have it open. But like you said, it's only useful if you have the expertise to tell what's good from what's bad. You talked about content briefs. I have a built-in content brief skill where the human-in-the-loop flow is important. If I ask it to create a brief, it uses the skill I've built, which includes the keyword, customer journey stage, search intent, target ICP, and SERP analysis. But it also asks me specifically whether I have a unique angle, or whether I've interviewed a subject matter expert I want to add. Then it runs based on that. So that human-in-the-loop step is really important, like you mentioned.
Kateryna (18:57): Yes. In our agency we have a bunch of skills, including a brief skill. It's one of the things that makes work fun, because a skill is something you can reuse, improve, and upgrade. That's the most exciting thing in Claude.
Usama Khan (19:30): Yes. That's how you really get value from it. If you're repeating a task weekly, that's where AI is most useful. If you set it up properly and keep improving it, you extract the most benefit.
Kateryna (19:30): Yeah.
Usama Khan (19:30): Kate, producing hyper-specific content at scale requires writers who can extract and translate expert knowledge. That's a rare skill. Most content writers are trained to research a topic on Google and summarize what already exists, which produces generic content. You've built a great team. How do you find and train writers who can do the opposite and produce differentiated content?
Kateryna (20:00): This is the hardest question for a content agency. The people I work with, we've been together for a few years. I haven't hired anybody recently, though I'll have to start eventually. It's the hardest thing in all my years running a business.
The people you find are not trained. This is my view, maybe a bit biased, but I think SEO culture has had a bad impact on the writer market. When you find a writer who was writing before AI appeared, the way they wrote and the way they write now isn't the level of content I want to publish. They couldn't do it well before AI. They do it a bit better now with AI, because it makes their writing more polished, but there are still no unique thoughts.
There's a popular opinion that AI has worsened content. Actually it's improved content, because if you look at what writers produced before, it was way worse. Most writers weren't great, and great writers are always rare. My team weren't all perfect writers originally. But over time, when you work with them and educate them on how the work needs to be done, they improve a lot. My team has upgraded itself enormously over the years. It's always hard, though, because it's very time-consuming, especially for me. We even have an in-house editor, but I still need to control the quality of some pieces all the time.
Usama Khan (22:09): That makes sense. So it's about spotting the baseline skill at your agency, then working with them to continuously improve, so you can create better content for your clients.
Kateryna (22:23): Exactly. Training people is a very difficult thing in general. I've been thinking about new writers who just entered the market. They'll use AI, but they don't have the experience of doing it without AI, and that's going to be a challenge. You can't really be a junior in this AI era. AI helps you level up to mid-level or even senior pretty fast, depending on how you define those levels. But that's the problem. When you think about new jobs and how people without experience are going to get hired, they'll perform much worse than people who have pre-AI experience and are now performing much better with AI. Comparing newcomers to them, it's a big question for now.
Usama Khan (23:24): This tracks back to a saying I keep seeing on LinkedIn: AI accelerates what you already are. If you're a skilled, great writer, you can do more with AI while keeping the same quality. If you're generic or not a great writer, you'll accelerate AI slop.
Kateryna (23:47): That's the problem, yes.
Usama Khan (23:48): Every company has AI tools now, and volume is basically free. So what's the new bare minimum for content that's actually differentiated, and how can teams hit that bar at reasonable scale when client budgets are shrinking and expectations keep growing?
Kateryna (24:07): You need a strategy, which you could also call a minimum viable strategy. There's a common belief that to perform well in AI search you need a lot of content. You need content, but whether you need a lot of it is a real question. You might need much less than you think.
To get results without investing a lot, you need to be very specific about what you publish and how you'll distribute it. If you don't have an answer for how you'll distribute it, maybe it's better not to write it at all. You need a well-weighted balance between what you can create and what can generate results. It can be very few pieces at the start, but very well positioned and unique. You analyze what competitors write and figure out what uniqueness you can bring.
You also have to understand the trade-offs. If you want to invest in a report and it takes more than three months, that's not something that shows results right away. Maybe you start with something less time-consuming. But I would never suggest covering as many keywords as possible just hoping one ranks. That's not how it works anymore. It didn't even work before, unless you had big link-building budgets. Now it matters even more to be specific and limited.
Usama Khan (25:51): Do you primarily focus on bottom-of-funnel content, or do you also write top and middle of funnel?
Kateryna (26:01): We don't really have that division. We just know what we need to rank for. We know the position and the audience. The priority is always bottom of funnel. That's where we start. It's an inverted pyramid. Top-of-funnel content can happen too, depending on how you write it.
Normally, educational content isn't something people need anymore, because they get that education from AI chat, or from Google's AI answers. But you can still create educational content with your specific frameworks, just to explain your approach. Or you can create controversial pieces. The market thinks one thing, and you explain why you think differently. That's an important piece of content for building brand authority, because it differentiates your brand. You say something nobody else is saying. That can count as top of funnel, but you still need it for positioning. So we create that too. But when we start, we want results as fast as possible, so we begin at the bottom.
Usama Khan (27:14): That makes sense. A lot of people discard top-of-funnel content, and I understand why. You hit it: it's usually a good idea to start with bottom of funnel, because that's where you see the highest ROI. But I still suggest clients start at the bottom and then move up. If you're a new category, top-of-funnel content still helps educate the reader, especially if you have genuine frameworks that help your ICP. And if you have strong distribution, a strong newsletter, that's a good channel for top-of-funnel content too.
Kateryna (27:56): Yes, exactly. Distribution is very important. Right now we even have content packages with distribution built in. The more you appear on other websites, on LinkedIn, on YouTube, on your own owned media, the more results you get.
Usama Khan (28:14): Absolutely. That's a wrap, Kate. If you got value from this, subscribe wherever you're listening, and I'll see you on the next one.
Kateryna (28:23): Thank you.
