A few weeks ago, I jumped on a call with Tal Klein, CMO at Lakeside Software.
Working as an engineer, stumbling into technical marketing at NetScaler, and building marketing functions at Citrix and Bromium has given him unique insight into what’s working – and what isn’t – in B2B marketing today. So, that’s what we spoke about.
In this conversation, we cover AI’s real impact on buying behavior, why he keeps LLMs out of the middle of his content process, and the one uncomfortable truth about the CMO role.
You didn’t start in marketing. How did you end up here?
I fell into it, honestly. I started as an engineer, but I was always put in front of customers or non-technical teams in the organization to explain what we were building.
If someone needed to understand why they needed another DLT tape drive, they’d send me. I was good at translating technical detail and nuance into metaphors that non-engineers could understand. Pretty soon, I realized that I was better at that than I was at being an engineer.
When NetScaler came calling with a marketing role, I was kind of offended. I was like, “no man, I’m an engineer.” But they managed to convince me.
NetScaler told me that their buyers were technical people, like me. They don’t want marketing speak, they don’t want to just hear that we have the best load balancer. They want to know the outcomes it drives – how does it do caching, SSL offload, and how does it do it better than other products?
From there, a great mentor at Citrix named Wes Wasson – who is sadly no longer with us – helped me see I had two paths: go deeper into product, or round out my skills across marketing.
He pushed me toward the broader path, so I spent two years wearing every hat at Citrix. Director of marketing, then senior director, and so on. Each step built on the technical foundation I already had. That’s how I eventually got my first CMO job at Adallom.
What’s the biggest shift you’ve seen in how buyers evaluate vendors over the past five years?
I hate to say AI, but it’s AI. It’s shifted how everybody makes decisions.
The traditional enterprise buying cycle was pretty predictable. You talk to peers, evaluate a few vendors, run a bake-off, go to procurement, procurement pulls the Gartner Magic Quadrant, identifies three more vendors, and the whole thing repeats.
Basically, all the research was done by humans. You’d Google, download whitepapers, poke around a vendor’s site. That’s just not the case anymore.
Over 50% of the traffic on Lakeside’s site is LLM traffic. ChatGPT and Claude are reading websites, not humans. It’s hard to overstate how many of the assumptions we used to rely on for marketing that breaks.
Intent signals, for example, used to mean something. Five people from AcmeCo browsing your pricing page was a signal worth acting on. Now it just tells you an LLM stopped by. You don’t know what question triggered it, who asked, or what conclusion it reached. The buyer has formed an opinion before they ever talk to you.
So how do you adapt? What’s working right now?
You’ve kind of got to design your site for both a real person and their digital twin.
Not in the sense of building two separate sites, but in how you think about it. Assume your ideal customer also has an LLM acting on their behalf, so your content needs to answer the questions that both the human and the AI are likely to ask.
For LLMs, that means structuring content in a way that’s easily parsed and synthesized. Think hero’s journey or Pixar story structure: small, meaningful pieces building along a clear narrative arc. LLMs consume coherent stories well and struggle to extract meaning from dense, disorganized content.
For humans, it means getting right to the point and offering highly tactile experiences from the off. By the time a real person lands on your site, they’ve likely been pre-educated and already formed a view. Skip the lengthy explanation.
Beyond the website, we’re finding real dividends in places that might seem small. Engaging in forum threads, niche subreddits, even responding to old tweets. LLMs visit those communities regularly and learn from them. A genuinely useful response in the right forum doesn’t just reach that community. It potentially shapes what an LLM says when someone asks about your space six months from now.
How is your team using AI to create content, and how do you stop it sounding like everyone else?
Our approach is simple: AI at the beginning, AI at the end, but humans all the way through the middle.
At the ideation stage, AI is genuinely useful. First drafts, outlines, getting ideas on paper fast. At the copy-editing stage, it helps catch errors and tighten language. But for the actual content creation, it’s 10% AI or less, and that’s intentional.
Our audience at Lakeside Software is highly technical IT professionals with a finely tuned sixth sense for inauthenticity. If something sounds assembled rather than thought through, they notice immediately. Even with extremely detailed prompts, current LLMs can’t reliably capture brand voice, let alone the depth of perspective that resonates with technical buyers.
So often, I see great ideas over-polished by AI. Maybe people are losing trust in their voice. It’s okay not to know exactly the right word at the right time; it’s not okay to wash out genuinely good ideas by using AI to turn one sentence into a paragraph.
What can you genuinely tie to revenue, and what are you still funding without clear attribution?
I’m lucky to work with investors at Insight Partners who are specific about what they care about.
Marketing’s impact on revenue is undoubtedly our core KPI – are we generating leads that convert to opportunities, that convert into revenue? And are we doing it at a conversion rate and velocity that either meets or exceeds our targets?
Within that, how I allocate budget is largely up to me. For example, over my career, I’ve found that I’ve spent an oversized amount of my budget on sales enablement. Marketing can generate excellent, highly relevant leads. But if the sales team can’t articulate the same value propositions in the same language, you’ve built a disconnect that quietly kills pipeline. You’ve advertised cruises and your salespeople are selling theme park tickets. Velocity is the metric that surfaces this. If marketing-generated leads are converting but stalling, the handoff is usually broken rather than the lead quality.
I treat sales enablement as part of the marketing investment, not a separate line item. You need the whole system to work. The board doesn’t penalise us for spending on things that aren’t immediately measurable. They care about the outcome more than the clockwork.
What’s your top budget priority overall?
Partner marketing, which reflects how Lakeside goes to market. We rely heavily on channel partners, ISV alliances, and GSIs. But the principle applies broadly.
Partners can consume an enormous amount of your time if they’re not equipped to bring you the right leads. A good partner will surface every opportunity that fits your category, and so if you’re not precise, you risk spending all your capacity on POCs you’ll never win.
In general, you can’t say no to channel partners without damaging the relationship, so you must invest upfront in tuning the joint go-to-market. That means making sure they understand what a good-fit opportunity looks like, what the differentiated use case is, and what message to lead with.
I learned this the hard way at Bromium. We offered a kind of advanced endpoint protection solution. Our partners kept bring us leads that didn’t really fit, like opportunities to displace McAfee, Palo Alto, things like that. But really, our product wasn’t a replacement for endpoint protection, it was a safer way to browse the web. That’s kind of a different pitch – our selling point was that organizations could browse without fear of compromise via browser-based attacks. But out partners didn’t get that. Anyway, my advice is to help your partners understand your product’s actual use case- or they’ll bring you the wrong fights.
How do you bridge the gap between marketing and engineering or product?
Honestly, I don’t think you can do it well without domain expertise. I won’t hire someone in marketing who doesn’t have a foundational understanding of what products we make, how customers use them, and why they buy them. That’s a hard line.
The core requirement is trust, and trust between marketing and engineering doesn’t come naturally. Engineers and product teams are often sceptical of marketing because they’ve seen marketing make promises the company can’t keep. The only way to build genuine credibility is transparency about what marketing is doing and why.
That means being clear. Explain how you’ll educate the market on what you’ve built, tease roadmap items so buyers are ready when you ship, and keep sales focused on what’s available today while building pipeline for what’s coming. Ensure you keep engineers informed about what’s trending so they don’t burn cycles on something the market is moving away from.
Over-communicate on clarity. Before any cross-functional initiative, make sure everyone at the table genuinely agrees on what you’re there to accomplish. Eliminate the daylight early, because small assumptions at the start become expensive misalignments six months in.
You mentioned speaking the buyer’s language. What does that look like in practice?
Your brand voice should never be in a different vernacular than your ICP. It sounds obvious, but it’s violated constantly, especially at conferences.
Take the ServiceNow Knowledge Conference. Walk the floor and you’ll see booths everywhere with messaging like “faster ticket resolution” or “better end-user experience.” Those outcomes are real, but the people attending that conference don’t think in those terms. They think in MTTR, first correct assignment rates, ticket reassignment ratios. Those are the KPIs they’re measured on, and those are the words that actually stop them in their tracks.
If you want someone who lives and breathes CMDB to pay attention, don’t talk about the entire universe of ITSM. Get specific. Show them you understand their exact problem in their exact language. That’s what creates genuine engagement, not polished generalities that say everything and mean nothing.
The same principle applies when you shift context. At a business strategy conference, drop the acronyms entirely and draw a clear line from your product to bottom-line business impact. Your ICP is rarely one person. There’s the economic buyer, the technical evaluator, the CFO signing off. Each has a different frame of reference. Speak each of those vernaculars individually, because there is no single message that magically resonates with all of them.
What’s one uncomfortable truth about marketing that most leaders won’t say out loud?
The CMO is a dinosaur.
I say that as a CMO who’s proud that many people who’ve worked for me are now CMOs themselves. But the traditional CMO role is being absorbed by other parts of the organisation, and in many cases that’s the right outcome.
If I were founding a company today, I’d think hard about whether I actually need a CMO, or whether I need a strong demand gen person embedded in growth, a strong product marketer embedded in product, and a strong sales enablement person embedded in sales.
Most modern founders and CEOs can already articulate their vision and narrative. That used to be the CMO’s most important job. If you hire a CMO and you’re already that type of leader, you’re often just introducing friction. Your message must travel through an extra person before it reaches the market.
The specific disciplines haven’t gone anywhere. ICP definition, lead scoring, PR, brand awareness: these are increasingly specialised functions that sit naturally elsewhere in the org, supported by specialist agencies where needed.
The role that still needs an experienced generalist at the top is a company where the CEO isn’t a natural storyteller, or where the marketing function is complex enough to require coordination across multiple buyers, geographies, and product lines. That’s not every company. Being honest about when you need that person, and when you don’t, is worth more than hiring for the title.
Tal Klein is CMO at Lakeside Software. He writes regularly about AI, enterprise technology, and marketing strategy on LinkedIn.
ABOUT BORA
We’re Bora. We work with security companies to turn complex technical capabilities into clear, credible market narratives.
Get in touch for a free 30 minute consultation. If we’re not the right fit, we’ll help find someone who is.
