Clay’s Everett Berry on AI and new shape of go-to-market

How is AI already changing go-to-market strategy, operations, and leadership? In this interview, Everett Berry, Head of GTM Engineering at Clay, offers a front-line view of what leading companies do differently, where revenue growth is likely to shift next, what distinguishes scalable AI adoption from stalled pilots, and how CEOs and boards should think about revenue leadership in a period of rapid change.

Everett Berry is Head of GTM Engineering at Clay, an AI-native go-to-market platform. He began his career as a software engineer, before moving into go-to-market, where he has spent the past several years helping revenue leaders build scalable systems for data, automation, targeting, and execution.

You’re the Head of GTM Engineering at Clay, an AI-native go-to-market platform, which means you work closely with revenue teams as they turn AI into repeatable go-to-market workflows. What are you seeing leading companies do differently as AI becomes part of how they run go-to-market? What changes most in how they grow when AI is truly embedded?

The biggest difference is that AI has dramatically increased the rate at which go-to-market teams can ship. The closest analogy is what has happened in engineering. Just as AI is enabling engineering teams to write more code and release software faster, it is doing something similar for revenue teams in the form of plays. Teams can launch new campaigns faster, test new messaging faster, enrich their data faster, and build automations that improve sales productivity without the same level of manual coordination that used to slow everything down.

That alone is meaningful, but the more important shift is what that speed unlocks. In the past, many go-to-market initiatives took so long to align and launch that organizations were naturally conservative. If a campaign required months of planning, cross-functional approvals, and operational work, there was pressure to get it “right” before it went live. AI changes that equation. It becomes much easier to test multiple ideas, compare messages, try new segments, and refine what works in real time rather than treating every launch as a major bet.

That increase in experimentation is where the strongest companies are separating themselves. They are not assuming they can predict in advance which campaign, persona, or message will land best. Instead, they are building an operating rhythm around testing and feedback. AI lets them run more experiments, gather signal faster, and scale what works with less friction. In practice, that means growth becomes less about finding one perfect motion and more about building a machine that can learn faster than competitors.

It also tends to favor a more centralized go-to-market model. When the data, personas, workflows, and messaging live in a coordinated system, teams can move quickly without creating chaos. You can test more because the inputs are cleaner and the feedback loop is tighter. That is where AI starts to become truly embedded. It is not just sitting at the edge of the workflow helping someone draft an email. It is part of the infrastructure that helps the organization decide what to do next.

In the next two to three years, where do you expect the biggest shifts in revenue growth to show up based on what you’re seeing in development right now?

I think those shifts will show up in three places: org design, buying motion, and product model.

On org design, I can see two very different but equally plausible futures. One is simplification. In that world, organizations converge around a smaller set of roles. You have go-to-market engineers or highly technical operators who build automations, manage systems, and scale what is working. Then you have sales reps whose job is focused much more on direct human intervention: stepping into the moments where judgment, trust, and relationship still matter most. In that model, the system does the heavy lifting around process, data, and orchestration, while the human seller applies expertise where they has the most leverage.

The other future is actually more specialized. Instead of fewer roles, you could see organizations break the revenue motion into narrower pieces. A seven-stage sales process might become seven distinct areas of expertise, with one person focused on proof of concept, another on legal and procurement, another on onboarding, and so on. On paper, that sounds cumbersome, but if the process around those people is well coordinated and AI handles much of the operational handoff, the buyer may actually get a better experience because they are engaging with a true specialist at each critical moment.

Whichever direction wins, I do think the profile of the sales rep is going to change. The most valuable people may be less like the classic all-purpose seller and more like highly knowledgeable experts in a particular industry, product, or use case. They may not even be elite sales operators in the traditional sense. What will matter is whether they can show up with credibility, answer detailed questions, and create trust. The surrounding automation can increasingly handle the research, data gathering, list building, and process support that used to consume so much of a rep’s time.

The buying motion itself is likely to split by customer type and complexity. At the low end, many customers will continue to buy through self-serve or product-led motions. Just above that, I think you will see more customers buy from agents, especially in SMB and simpler commercial use cases where buyers want answers but do not necessarily need a fully human sales process. At the higher end of the market, human reps will remain critical, but their role will shift. They will spend less time prospecting and more time building trust, shaping solutions, and staying close to the customer over a longer relationship arc.

That may even bring back a more account-centric model. Instead of slicing the customer journey across separate pre-sales and post-sales teams, some companies may decide it is better to keep one person or one small team close to the account over time. In a strange way, AI may automate much of what SaaS sales professionalized over the last decade and push human sellers back toward deeper relationship work.

Then there is the product side. Companies whose products are essentially UI layers on top of workflows will increasingly have to ask whether they are built for an agent-driven world. If software is going to be used not only by humans but also by agents, then products and pricing models need to evolve accordingly. That likely means more API-first design, more consumption-based logic, and more thinking about which pieces of the product can be used as primitives inside automated systems. The companies that adapt to that shift will be better positioned for growth than those that remain built around a purely seat-based, human-interface model.

When you look at the most forward-leaning revenue organizations today, what separates the teams that successfully scale AI from the ones that stay stuck in pilots? What did they change in their operating model, whether in people, process, data, or governance?

The biggest difference is that the teams who succeed expect the early version to fail, or at least to fall short. They understand from the start that implementing AI is not usually a plug-and-play exercise. With traditional software, the organization often adapts itself to the product. With AI, it is often the reverse. The system has to be shaped around how the company actually works, how its data is structured, what success looks like, and where human judgment needs to stay involved. That makes the process messier, more iterative, and sometimes frustrating at the beginning.

The organizations that get stuck in pilots often misread that messiness as proof that the approach does not work. They try a first version, the ROI is not immediately obvious, the metrics are uneven, and the initiative stalls. The more effective teams treat that phase as normal. They give themselves enough room to refine the use case, clean up the data, improve the prompts, adjust the workflow, and keep going until the value becomes clear. In many cases, it is not the first or second version that changes the organization. It is the fourth version that finally clicks and becomes something people rally around.

That requires leadership commitment, but it also requires a different operating model. The strongest teams create faster iteration loops, centralize more of the experimentation, and take data quality seriously. A lot of AI value depends on underlying data being usable, structured, and trustworthy. If that foundation is weak, the organization will struggle to get consistent results no matter how promising the tool looks in a demo. So in practice, scaling AI often means investing not just in a new application but in the core discipline of how information is managed and activated.

A useful example is Intercom and the way it has built around Fin. What stands out there is not just that the company launched an AI product, but that it reoriented both product and go-to-market around the opportunity. Using AI-driven targeting, the team was able to identify micro-segments that were especially strong fits. In Finn’s case, those ideal customers tended to have both a well-developed support knowledge base and a high volume of support requests. Once that profile was clear, the team could isolate and pursue specific segments where the product was especially likely to perform well.

That is the kind of operating discipline that separates scalable AI efforts from scattered pilots. The winning teams do not stop at the broad use case. They refine, segment, test, and sharpen until they find where the motion really works. Then they scale that. Many organizations never get that far because they give up in the awkward middle, before the system has had time to mature.

What’s a common, but now dated, assumption about how revenue work has to be done? What do you wish leaders understood about what AI can already do today, based on what you’re seeing in the field?

One of the most dated assumptions is that prospecting is fundamentally powered by individual rep magic. A lot of companies still operate as though the core of outbound success lies in each rep independently building lists, researching accounts, finding contacts, crafting messaging, and executing all of that with consistent quality. In reality, that model has always been far less reliable than organizations wanted to believe.

What matters much more is the quality of targeting and the quality of the data. At Clay, there is a phrase that captures this well: the list is the message. If you know which accounts matter, which people are relevant, what signals suggest timing, and what context should shape outreach, you are already most of the way toward a much stronger prospecting motion. AI is increasingly very good at that layer of the work. It can help identify the right companies, surface the right contacts, organize the right signals, and even generate sensible starting points for messaging at a level of speed and consistency that is difficult to match manually.

That does not mean human sellers no longer matter. It means the center of gravity shifts. The differentiator is less about who can spend three hours building a spreadsheet of accounts and more about who can step into a live conversation and create real value. The best reps are the ones who can be knowledgeable, credible, and useful in front of the customer. They are the ones who can ask smart questions, navigate nuance, and build trust. AI can remove a great deal of the repetitive work that used to sit upstream of that moment.

It also means companies should rethink how much prospecting activity really needs to be decentralized. For many organizations, especially those with clear ICPs and repeatable motions, centralizing more of the targeting, contact identification, and message development can raise the quality bar across the whole revenue team. Instead of asking fifty reps to each perform the same research and targeting tasks at different quality levels, a company can build that centrally, automate much of it, and let reps focus more of their day on the human work that cannot be fully automated.

That breaks with a lot of old assumptions about sales, including the idea that the best way to improve pipeline generation is simply to hire more reps and hope enough of them are exceptional. Increasingly, a stronger answer may be to build a better system, so that performance depends less on random variation and more on shared intelligence.

If you could give one message to CEOs and boards about how AI is changing revenue leadership, what would it be?

The message would be that the best revenue leaders now need to think like systems leaders. The strongest CROs I see are not just managing sales teams or chasing quotas. They are looking at go-to-market almost the way a product leader or engineering leader would look at a system they are responsible for improving. They think about the tooling, the data flows, the feedback loops, the campaign velocity, the points of automation, and the places where humans create the most differentiated value.

That is a meaningful shift in leadership posture. A modern revenue leader should have a point of view on where AI belongs in the organization, what work can be automated today, what should remain human-led, how reps are currently using AI in fragmented ways, and what needs to be centralized so the company is not just getting random local productivity gains. The question is not whether people are experimenting with these tools. In most organizations, they already are. The question is whether leadership is turning those scattered experiments into a coherent operating model.

There is also a talent implication that boards and CEOs need to take seriously. If AI can automate a meaningful share of repetitive revenue work, then the composition of the team will change. Some roles will shrink. Others will become more technical. Some organizations may need fewer traditional reps and more product experts, operators, or domain specialists. The companies that handle that shift well will not be the ones that simply cut headcount fastest. They will be the ones that redesign the work thoughtfully and align talent to the new model.

That is why I think the most useful framing is to treat go-to-market as a product or system. It helps leaders make better decisions in the present, and it also hedges against uncertainty about where the models go next. No one knows exactly how far agent capability will advance or how quickly. But leaders who are already building a modular, observable, improvable revenue system will be in a much stronger position if those capabilities improve rapidly. They will be ready to adapt, because they are already thinking in terms of architecture, not just headcount and process.

In that sense, AI is not simply changing how revenue teams execute. It is changing what revenue leadership is. The old model rewarded managers who could drive activity inside a familiar sales machine. The emerging model rewards leaders who can redesign the machine itself.