Artificial Intelligence has become one of the most discussed forces reshaping sales and marketing. Every week, new tools promise to automate outreach, generate leads, and replace manual workflows. But behind the buzzwords, what is actually changing inside go-to-market (GTM) teams? To better understand where things are heading, I spoke with Yoni Tserruya, co-founder and CEO of Lusha, about the evolving role of data, the rise of signal-driven sales, and why intelligence – not automation – is the most valuable layer in the modern GTM stack. Yoni Tserruya, Yoni Tserruya, Lusha Lusha The Shift From Sales Reps to Sales Infrastructure AI has dominated the sales conversation for the past few years. Where do you actually see it genuinely change GTM and not just create hype? AI is changing many aspects of the industry. One of the biggest shifts is moving from human-driven sales work — sales reps doing everything manually — to more agentic workflows. AI changes how you consume information, how you act on that information, and how the sales process itself works. But there's another interesting shift that's organizational. In the past, salespeople were the central persona because they generated revenue directly. Now the focus is shifting toward operations — revenue operations, sales operations, and similar roles. Those teams are the ones building the workflows and infrastructure that AI runs on. They have a lot on their plate to create growth. The truth is that everyone is still learning. No one has the complete playbook yet. Every company is figuring it out as they go. It's like everyone is writing their own playbook right now. You built a B2B sales platform without having a traditional sales background. How much of that outsider perspective was an advantage, and where was it a challenge? My background is mainly in data and product. I've always been interested in building systems that help people make better decisions using data. There was a big need in the go-to-market and sales world, so we built a product for it – Lusha. When developing it, we were focused on two main areas: the data layer and the sales feature layer. Looking back, our success clearly came from the data side. Where data itself was the product, we succeeded. Where we tried to compete on traditional sales features, we didn't do as well. That realization pushed us to focus fully on intelligence and data. That's where our strongest asset is and where we believe the industry is going. We're investing most of our effort in building the best intelligence platform we can, while assuming the sales features layer will be handled by other companies. Why "Spray-and-Pray" Is Fading What problems are GTM teams actually struggling with today? At the end of the day, GTM teams have one core job: build pipelines and bring in customers. For many years, the strategy was simple – send emails or make calls to as many people as possible and hope something works. That spray-and-pray model is becoming less effective. It's not even effective with automation anymore. The industry is moving toward relevant outreach, meaning you contact companies when they're most likely to want to hear from you. And there are a lot of signals that can indicate when that moment is. For example: a company is hiring a lot of salespeople, a new sales leader just joined, the company recently raised funding, website traffic is increasing, or they've just replaced their CRM. When you combine these signals, the probability of having a meaningful conversation becomes much higher than if you just reached out randomly. That's where the industry is going – and that's what we're trying to enable. Have you seen measurable results when your customers start operating this way? Yes. Companies that use signals instead of random outreach to their ICP audience typically see two to four times improvements in conversion rates. Interestingly, the volume of leads often goes down, but the relevance and accuracy of each one increase significantly. What we're seeing now is a deliberate shift toward fewer leads but deeper engagement with each one. What we're seeing now is a deliberate shift toward fewer leads but deeper engagement with each one. What we're seeing now is a deliberate shift toward fewer leads but deeper engagement with each one. Companies want fewer leads but are spending more energy on each lead. That's where efficiency really comes from. There's a lot of talk about AI systems failing not because the models are bad, but because the underlying data is poor. How do you approach that? The magic in the AI world happens when you provide the most context to the AI. You can use a platform without connecting any of your internal systems, and over time, it learns your behavior and improves recommendations. The more you use it, the better we understand your ICP and the better our recommendations become. But when you connect systems like your CRM, the results improve dramatically. Your CRM contains information about closed-won and closed-lost deals, which gives the AI a strong signal about what your ideal customers actually look like. We have models that align with your recent pipeline history in our data and start predicting which types of companies are more likely to close in the future. Those models are already working pretty well. So the more context you give the platform, the better the outcome will be. CRM is one of the best sources for that because it's the system of record. "Sales Streaming": Borrowing a Page From Spotify You've coined the term "sales streaming" inspired by Spotify's recommendation engine. But Spotify still gets things wrong after years of refinement. What are the real limitations of applying that model to B2B sales? The vision came directly from Spotify. You pick one or two songs, and the system continues recommending similar music automatically. Our vision was the same for sales organizations: Once we understand your ICP and your best customers, we stream you similar companies so you don't have to search manually. That's the evolution we're imagining. Once we understand your ICP and your best customers, we stream you similar companies so you don't have to search manually. That's the evolution we're imagining. Once we understand your ICP and your best customers, we stream you similar companies so you don't have to search manually. That's the evolution we're imagining. Early versions of Lusha focused mostly on ICP similarity, and we did a pretty good job of surfacing similar companies. Today, the logic has improved significantly, as it now includes signals that make the recommendations much more relevant by providing timestamps. Instead of just finding companies that look like your customers, we can now identify companies where something meaningful just happened. That timing makes a big difference. We're working hard on it and investing heavily because we believe it will be the endgame eventually. Will AI Replace Sales and Marketing Jobs? The bigger question many people are asking: does AI replace sales and marketing roles, or transform them? I think it will eliminate a lot of existing jobs,  especially technical white-collar jobs that AI can perform better than humans. That's going to happen. But humans will still be essential for several things. Trust between people still matters enormously. Large deals require negotiation and relationship-building. Humans are also needed to design systems and define a strategy. AI can do the work, but someone behind it needs to think about how things need to work together. AI can do the work, but someone behind it needs to think about how things need to work together. AI can do the work, but someone behind it needs to think about how things need to work together. What's interesting is the inversion happening. Some of the highest-paid knowledge workers today may be the most exposed to disruption, because AI is smarter and faster at execution. Meanwhile, roles involving physical work, human interaction, and emotional intelligence may remain much more resilient. I also think we'll see more people move toward industries like entertainment, hospitality, and experiences – the things that require a human presence. It's moving fast, and you're always surprised by where it goes. But that's my perspective today. Intelligence as the Next SaaS Layer Where do you see the next phase for a platform like Lusha? Partnerships will play a major role. Our vision is to become the intelligence layer for other SaaS platforms. Many systems – CRMs, sales engagement tools, customer success platforms – provide strong functionality but limited intelligence. If you embed intelligence directly into those systems, they become far more useful without requiring additional effort from the user. We want to be the intelligence provider, working behind the scenes with zero maintenance on their end. That's the direction we're investing heavily in this year – becoming an intelligence-as-a-service for other SaaS platforms. The Builder Era of Go-to-Market More operators today want to design automated AI-driven workflows. Are you building with that audience in mind? Yes, and it's a major focus. Everything we build is API-first. Builders – anyone designing workflows or agents – can access our data through APIs and integrate it into whatever they're building. We also provide connectors to automation platforms such as Zapier, Make.com, and n8n, so teams can build workflows without starting from scratch. And within the platform itself, you can work agentically: upload any source, pull in our data, and get the output you need. We have a few layers, but all are designed for flexibility and tailored to each team's needs. From Product-Led Growth to Agent-Led Growth Five years ago, we were all talking about product-led growth as the new model. Now it's agent-led growth. What makes this moment different from previous waves? Agent-led growth is definitely the current wave. What makes this moment unique is that companies need to transform in two ways simultaneously. First, their products must incorporate AI. Second, their internal operations must also run on AI so they can build faster and operate more efficiently. First, their products must incorporate AI. Second, their internal operations must also run on AI so they can build faster and operate more efficiently. First, their products must incorporate AI. Second, their internal operations must also run on AI so they can build faster and operate more efficiently. Each change alone is significant. Together, they create a massive shift in how companies operate. You need to change your product and your culture at the same time, and that takes courage. A lot of people stay in their comfort zone, and that's just not a viable place to be right now. The crazy part is that every six months, there's a new wave to adopt. It was agentic first, then the focus shifted to builders, now it's AI-assisted coding and AI in support. Only the companies that move fast enough to keep up will make it. The Future of the GTM Stack Is there a world where all of this consolidates into one platform? Or does the stack stay fragmented? Execution tools will probably remain specialized – sending emails, running campaigns, those things will have dedicated platforms. But intelligence is different. The companies that win will be the ones that know what to do: which accounts to focus on, when to reach out, and why. The companies that win will be the ones that know what to do: which accounts to focus on, when to reach out, and why. The companies that win will be the ones that know what to do: which accounts to focus on, when to reach out, and why. Execution is relatively easy. Knowing where to focus your energy is the real competitive advantage. That's the layer we're focused on owning. The promise of AI in sales is often framed around automation – doing more, faster. But this conversation suggests the real shift may be happening somewhere else entirely: not in the volume of outreach, but in the quality of the moment you choose to reach out. The real question isn't whether your team can reach more people; it's whether you'll know the right person, at the right moment, before your competitor does.

Internet Computer

Lusha

Beyond the Perimeter: Securing AI for the Quantum Era

Read me on HackerNoon

Nominated for 2022 - HackerNoon Contributor of the Year - Robotics

Nominated for 2022 - HackerNoon Contributor of the Year - Founder Interview

Nominated for 2022 - HackerNoon Contributor of the Year - Hacking

Nominated for 2022 - HackerNoon Contributor of the Year - Innovation

Portfolio

Too Long; Didn't Read

The End of Spray-and-Pray: How Intelligence-as-a-Service Is Reshaping GTM

The End of Spray-and-Pray: How Intelligence-as-a-Service Is Reshaping GTM

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

A Female CTO Shares Her Thorny Journey In the Male-Dominated Tech Industry

Beyond the Perimeter: Securing AI for the Quantum Era

The Agentic AI Maturity Gap: Combining Orchestration, Observability, and Auditability

The 4 Marketing Channels That Took My Bootstrapped Startup to 1M Users

Adopting AI in Sales and Marketing

Building Smarter Chatbots: Enhancing AI Responsiveness to User Needs

A Female CTO Shares Her Thorny Journey In the Male-Dominated Tech Industry

Beyond the Perimeter: Securing AI for the Quantum Era

The Agentic AI Maturity Gap: Combining Orchestration, Observability, and Auditability

The 4 Marketing Channels That Took My Bootstrapped Startup to 1M Users

Adopting AI in Sales and Marketing

Building Smarter Chatbots: Enhancing AI Responsiveness to User Needs

Light-Mode

Classic

Newspaper

Minty

Dark-Mode

Neon Noir

Minty

HN StartUps