Data is commoditizing. More providers, more alternatives, more coverage. On paper, that should make the data layer less interesting. But Yoni Tserruya, CEO and co-founder at Lusha thinks the opposite is true.

When data was consumed by a human, a bad record meant a wasted call. When data is consumed by an AI agent making autonomous decisions at scale, a bad record compounds. The agent routes wrong, scores wrong, prioritizes wrong. And it does it a thousand times before anyone notices. The problem isn't scale. It's that agents don't catch their own mistakes.

That's what pushed Lusha to make a deliberate call: move away from the all-in-one platform, go deeper into the data layer. API-first. Open by design. Usage-based pricing. No seats required. Because the next consumer of their data isn't a salesperson opening a browser tab. It's an agent calling an API at scale, with no patience for stale records.

Lusha turns ten this year. It started as a side project, a bootstrapped tool for pulling contact data mostly for recruiters, until Google called because it wanted to buy licenses. That was a decade ago. And the way Yoni tells it, the next decade doesn't look anything like the last one. Because the consumer of data is completely changing.

Dr. Anthony Annunziata started at IBM building better memory and transistors out of magnetic materials. The work was good, but he kept feeling the same itch: get the thing out of the lab and into the world where it actually matters. That instinct shaped everything after, through quantum computing, AI for science, and now IBM's open-source AI strategy.

The physicist's habit is first-principles thinking, stripping something down to what truly makes it work. But research and business part ways at the definition of done. In research, you discover, publish, and move on. In business, a demo is nothing. It doesn't work until someone else uses it, trusts it, and pays for it. He sees the same gap with agents today: demoing one is trivial, shipping one to production is the hard part.

The throughline is where your edge actually comes from. Frontier models are extraordinary, but progress on them has slowed while the cost of pushing further climbs, so the model itself stops being the differentiator. What's left is yours to build: your data, an eval strategy grounded in your real use cases, and the choice to run smaller models on your own infrastructure rather than handing your proprietary edge to a tech company.

In 2011, three co-founders from Missouri walked into Y Combinator with a simple observation: every SaaS help forum had the same unanswered thread. "When will you integrate with X?" Nobody was building the infrastructure to close that loop. They did.

Paul Graham told Wade and every YC batch the same thing: be a cockroach. Don't raise money. Survive. Adapt to whatever comes next. Almost nobody listened. Zapier did. Fourteen years, $1.3M raised, bootstrapped to a $5B valuation. Every time a funding opportunity came up, they asked one question: what is actually holding us back right now? The answer was never the balance sheet.

Then the SaaSpocalypse arrived. SaaS stocks cratered. Every software company started asking: what part of what we built actually survives this? Wade's answer is specific. Not the 8,500 integrations. The governance layer, the auth infrastructure, the background automation reliability stack.

And then Wade, in this episode of Revenue Engine Masters podcast, says something incredible. They built their entire business on drag and drop. It made them. And they're already moving past it. Natural language builders outperform it on activation. They measured it. They called it "yap to zap" and didn't look back. For Wade, this adaptability applies to humans too, not just companies. The best operators today aren't the ones prompting AI and moving on. They're the ones staying in the loop, iterating on context, pushing until the output is actually good. That's the human role now. In the outer loop around it.

Paul Graham's advice was about never optimizing for the conditions of today at the expense of your ability to adapt to whatever comes after. Zapier understood that in 2011. They still do.

Barry Dauber's job running GenAI GTM at Databricks is to convince enterprises to move faster on AI. So when he sat down at a VC dinner last year and everyone went around the table sharing what kept them up at night, the room got a surprise when his answer was "AI hype".

Barry is not a skeptic. He believes AI is transforming everything, and he has the customer stories to back it up. But he has also watched enough enterprise pilots stall, enough Monday morning messages come in saying the data was not ready, to know that the gap between what AI promises and what organizations can actually deliver is real and wide.

In this episode, Barry and Elio work through what that gap looks like from the inside: why most companies are not as ready as they think, what the early GTM lessons from MosaicML taught him about building before the market understood what you were selling, and what it actually takes to move from experiment to production at scale.

Remy Piazza, Principal & CRO at Powerhouse Consulting, started his career at Xerox, the godfather of solution selling.

In this episode, we explore why despite AI and evolving tech stacks, selling remains fundamentally human.

From mastering discovery to building trust, Remy shares why curiosity and genuine connection will always outperform automation.

Elio interviews SK Ramakuru about what RevOps really is: the “engine” behind selling that turns chaotic people/systems/data into repeatable motion. SK shares his path from Hyderabad, India—where early exposure to big tech sparked his ambitions—into consulting, then a Master’s program in the U.S. that deepened his stats and data skills. An experience of temporarily stepping into a selling role, and struggling badly, cemented a core belief: selling is art, not science. Data can point you toward opportunity, but performance depends on preparation, context, relationships, and understanding customer pain.

SK explains his analytical approach: in RevOps, the goal isn’t perfect data—it’s decision-guiding data that drives behavior and feels fair, especially in territory planning. His mental model is: start by trusting nothing, validate assumptions, slice data from multiple angles, and build a clear story that leaders and the field can understand. As MongoDB scaled from ~600 to ~2,500 sellers, he emphasizes designing territories that match the right accounts to the right reps, remain equitable, require minimal additional headcount, and are explainable and auditable.

SK and Elio also talk about how AI is changing both how RevOps works and what signals matter. Traditional indicators like website visits and content downloads are less reliable as buyers use answer engines (ChatGPT/Gemini) instead. That forces new time- and revenue-based indicators and stronger foundational data orchestration. SK argues the market is flooded with “AI wrappers”; the winners will orchestrate messy data across systems, not just generate answers.

In this conversation, Alan Gonsenhauser, Founder & CEO of Demand Revenue, explains why most go-to-market engines don’t break in the salesroom — they break long before buyers ever talk to sales. With only about 5% of accounts truly in-market at any given time, companies must stop treating marketing as demand capture and start building brand equity as future pipeline. By the time buyers engage, most of the journey is already complete, and without brand trust and preference, you never make the shortlist.

Alan also highlights a fundamental shift in buyer psychology, from FOMO to FOMU — Fear of Messing Up. In today’s risk-averse environment, deals are lost less to price or features and more to decision paralysis, making it critical for GTM teams to focus on de-risking the choice, not just promoting the product.

Finally, he reframes the role of AI, noting that it excels at convergent thinking — summarizing the past and increasing productivity — but not at disruptive strategy. The real ROI comes from fixing dirty data, fragmented systems, and back-office workflows, where clean, trustworthy operations quietly reshape revenue performance at scale.

“If you don’t have normalization of the metadata across the application layer, you cannot create the right systems of record to drive your systems of action.”

In this episode of Revenue Engine Masters, I sat down with Cliff Simon, founder & CEO of Polaris and one of the most thoughtful operators in the RevOps ecosystem. We went deep on the big shift happening in go-to-market: AI is forcing companies to confront their foundations—metadata, systems of record, systems of action—because bad data now leads to sh*ttier outcomes, faster.

Cliff breaks down why AI readiness is no longer a RevOps-only problem, but a company-wide mandate increasingly driven by CIOs, CFOs, and cross-functional leaders trying to decide what to build, what to buy, and how to ensure their teams are actually enabled.We explored how Polaris is helping companies navigate this moment with two offerings—AI Readiness Audits and GTME-as-a-Service—and why the new generation of AI-native GTM operators needs both strategic oversight and architectural guardrails.

He also shares how to build frictionless B2B buying experiences, the same way consumers buy: with credibility, simplicity, and trust.We closed by discussing speed vs. quality in building AI systems, the importance of doing work manually before automating it, and how to think about the emerging role of autonomous workflows. A super relevant conversation for anyone building GTM teams in 2025.

‍

RevOps isn’t a help desk—it’s a product function.

“AI can provide the data, we can automate the workflow, but it can't design the experience that people actually want to use.”

Revenue operations is less about managing tools and more about designing experiences people actually want to use.

Mollie Bodensteiner, Senior VP of Operations at Engine and former operations leader at Deloitte, explains why most ops teams become their own bottlenecks. She says successful revenue operations requires treating internal stakeholders like customers and that most enterprises over-engineer solutions instead of learning fast.

She shares how Engine scaled 70% year-over-year by organizing ops into pod structures—eliminating meeting fatigue and ensuring enablement isn't the last to know. That taught her to prioritize transparency over false promises when managing competing demands.

Mollie also walks through her prioritization matrix for deciding what makes the cut, warning that ops leaders who spend months on PowerPoints instead of testing will lose. She recommends hiring for curiosity over hard skills and setting quick wins within the first two weeks of onboarding. Her biggest warning: perfection kills momentum.

She predicts AI will separate operators who understand experience design from those who just automate workflows, freeing revenue teams to focus on the necessary instead of the unnecessary work that exhausts analysts before insights happen.

Haris Odobasic, co-founder of Revenue Wizards and author of The RevOps Pendulum, uses Newton’s pendulum as the perfect metaphor for RevOps: when one sphere moves, it transfers energy across all the others — but only if everything is aligned. In the same way, Sales, Marketing, and Customer Success must operate on the same plane, connected through shared data, processes, and goals.