Milena Traikovich is a powerhouse in the world of demand generation, known for turning cold analytics into warm, high-quality lead pipelines that actually convert. With a sharp eye for performance optimization and a background rooted in data-driven strategy, she understands that in today’s AI-saturated market, simply having a “cool tool” isn’t enough to survive. As the underlying models of artificial intelligence become a commodity available to everyone, the real battle moves from the code to the strategy. She has seen firsthand how quickly a competitive edge can evaporate when a rival can replicate a feature over a single weekend. Milena joins us to discuss the fundamental shift from building features to building moats, exploring how businesses can secure true, uncopyable defensibility through the lens of a sophisticated framework that prioritizes market framing, operational scale, deep systemic integration, and the wisdom of customer signals.
The conversation centers on the fragility of current AI-native growth, where high acquisition is often undermined by rapid churn. We explore the transition from being a mere vendor to becoming a category-defining “State” that dictates the language of the industry. We also dive into the mechanics of “Scale” that go beyond simple size, the “System” level of integration that makes a product indispensable to a customer’s daily operations, and the critical importance of “Signal,” which allows a company to understand the underlying motivations of its user base. Finally, we look at how these four elements form a self-reinforcing loop that separates the enduring market leaders from the temporary flashes in the pan.
We are seeing a startling gap in the market right now where AI-native companies are struggling to keep their customers compared to traditional B2B software firms. Why is this retention gap so persistent, and what does it tell us about the current state of AI growth?
It is a visceral problem for many founders right now because they are essentially running a treadmill that is speeding up while the floor is getting slippery. According to recent research, the median net revenue retention for AI-native companies sits at a mere 48%, which is a staggering drop when you compare it to the 82% benchmark we see in the broader B2B SaaS world. This tells us that while these companies are excellent at acquisition—growing bigger by the day—they are also leaking value just as fast. The underlying issue is that growth alone doesn’t equal safety; if your primary value is a wrapper around a frontier model, your customers aren’t sticking around because they aren’t losing anything vital by leaving you. Defensibility isn’t about what you own in your codebase; it is about the tangible pain or loss a customer would feel if they tried to walk away, and right now, many AI tools are just too easy to replace.
When you talk about a company achieving a “State” of advantage, you often mention that the market begins to use their specific language. How does a company move from just selling a product to actually framing the entire category?
Achieving “State” is a psychological and linguistic victory where you define the problem so effectively that your competitors have to use your dictionary to explain what they do. Look at a company like ElevenLabs, which didn’t just build a better voice AI; they became the absolute reference point for the entire category. This positioning helped them secure a $500 million Series D at a massive $11 billion valuation because they didn’t just offer a feature—they offered a standard. The data reflects this dominance clearly, with reports showing that roughly 95% of first-time voice AI buyers entered the market through ElevenLabs in the three months leading into early 2026. When you reach that level, you aren’t just a choice on a list; you are the starting point for the customer’s journey, forcing every other player to argue against your framework before they can even begin their own sales pitch.
Many people confuse being large with having a “Scale” advantage, but you argue these are two very different things. What does true, defensible scale look like in a modern business model?
Scale is only a true advantage if every new customer and every new transaction makes the business harder to beat, rather than just more expensive to manage. A perfect example of this is Ramp, which serves more than 70,000 organizations and processes a staggering $200 billion in annualized purchase volume. They don’t just sit on that data; they turn it into a competitive weapon through aggregated spend intelligence and AI-driven anomaly detection. By May of last year, their median customer was saving 50% more money and 32% more hours compared to just a year prior. That is Scale in action: the sheer volume of activity creates a feedback loop where the product actually gets smarter and more efficient the more people use it. It’s not just about having the most users; it’s about turning the accumulated activity into superior operating outcomes that a smaller competitor simply cannot replicate.
The concept of a “System” advantage seems to be about creating friction for a customer who might want to leave. How can a company embed itself so deeply into a client’s operations that moving becomes an organizational crisis?
System advantage is about becoming the nervous system of the client’s business, where your product is so woven into their operating model that removing it would require a total reorganization of how they function. Think about Vertiv, which integrates critical power and cooling equipment with predictive analytics and remote monitoring services. When a company uses Vertiv, they aren’t just buying hardware; they are adopting a lifecycle service that reduces the burden on their own internal operations teams. If that customer wanted to switch, they wouldn’t just be swapping a machine; they would have to rebuild their entire maintenance, risk management, and infrastructure-management model from scratch. This level of embeddedness creates a “sticky” relationship because the product has moved from being a tool used by the staff to being a capability that the company no longer needs to staff internally.
You’ve mentioned that “Signal” is about knowing the “why” behind customer actions rather than just the “what.” How does a company cultivate this kind of insight, and why is it so hard for competitors to copy?
Signal is the bridge between a raw dashboard and a strategic decision, and it requires a depth of data that takes years to accumulate. A company like Tempus is a master of this, linking molecular data—including DNA, RNA, and liquid biopsy results—with longitudinal patient records to help pharmaceutical teams. While a competitor might see that a drug is being prescribed, Tempus understands the “why” by analyzing real-world evidence and discovering biomarkers that support clinical trial design. This isn’t just about having a high volume of data; it’s about the proprietary ability to explain the meaning behind that data. Scale might give you the raw numbers, but Signal is the specialized knowledge that tells you exactly where you are winning and where you need to double down, and that kind of institutional wisdom cannot be replicated by a weekend software update.
How do these four pillars—State, Scale, System, and Signal—actually evolve as a company grows from a small startup into a massive enterprise?
The beauty of this framework is that it’s a continuous, self-reinforcing loop where each stage feeds the next. A startup usually begins with “State” because they lack the data or capital of an incumbent, so they win by naming a problem and framing the category in a new way. As they gain traction, that framing accelerates “Scale” by lowering acquisition costs and bringing in more users. That usage then creates the evidence and the deep integrations needed to build a “System” where the customer organizes their work around the product. Finally, all that activity generates the proprietary data that strengthens “Signal,” which in turn shows the company exactly how to sharpen their “State” and start the loop all over again. By the time a company reaches the enterprise stage, they have years of this loop turning behind them, creating a massive barrier that a competitor can’t just leap over by copying a single feature or lowering their price.
What is your forecast for the future of AI-native companies that fail to move beyond their initial technological novelty?
My forecast is that we will see a massive consolidation or “quiet exit” phase for AI companies that rely solely on model-driven features without building these structural advantages. Within the next twenty-four months, the novelty of “generative” capabilities will completely fade into the background, and we will see a return to the fundamental laws of business physics. Those companies that haven’t shifted their focus toward building a category-defining “State” or an embedded “System” will find their net revenue retention continuing to hover around that dangerous 48% mark until their acquisition costs eventually bankrupt them. The winners won’t be the ones with the most sophisticated prompts; they will be the ones who successfully turn their early AI-driven growth into a self-sustaining loop of data, integration, and market authority that simply cannot be copied by the weekend.
