11 min read
There's a term gaining serious traction in revenue circles right now: go-to-market engineering.
And if you've been hearing it more frequently but aren't entirely sure what it means or where it came from, you're not alone. The concept is still relatively new, and the role itself is still taking shape across organizations.
But make no mistake, it's already reshaping how the fastest-growing companies build their revenue functions.
I was one of the first go-to-market engineers at Clay, so I've had a front-row seat to how this has evolved. Here's what I've seen, what I believe is driving this shift, and where I think it's all heading.
What go-to-market engineering actually is
Let's start with the definition, because there's still a lot of confusion around the term.
Go-to-market engineering is the practice of applying engineering principles, automation, and AI to sales and marketing processes to build scalable, data-driven revenue systems. That's the cleanest way I can put it.
It's about taking the kind of systematic, iterative thinking you'd find in an engineering team and applying it to the problems that sales and marketing teams face every day: prospecting, outreach, personalization, segmentation, campaign execution.
The goal is to build systems that scale without simply throwing more headcount at the problem.
But we didn't always have this term or this role. Before go-to-market engineering became a thing, there were people doing adjacent work under different titles. Growth engineers. Revenue operations analysts. Marketing operations engineers. Automation engineers.
These folks were building workflows, connecting tools, and trying to automate the more manual, repetitive parts of the sales and marketing process. They were the tinkerers. The ones who figured out that you could stitch together a Zapier workflow with a Google Apps Script and save your SDR team hours of manual research every week.
The pre-go-to-market engineering stack looked something like this: traditional data sources and CRMs like Marketo, HubSpot, and Salesforce; orchestration tools like Zapier for no-code automation, or Replit and Google Apps Script for more complex logic; and then output destinations like Google Sheets or modern sequencers like Instantly or SmartLead.
It worked, but it was fragmented and required a lot of manual intervention to keep everything running.
Where the term actually came from
Like a lot of great ideas, this one started in a Slack channel.
When Clay's co-founders Karim and Varun were building out Clay's first go-to-market team, they were trying to put a name to something they were observing in their earliest customers. They needed a term that captured what these people were actually doing, because the existing labels didn't quite fit.
For context, Clay originally launched in 2017 and iterated over time. By around 2022, it had found a niche as a no-code workflow automation and list-building tool. The two main things it did well were helping teams find or import lists of contacts and then enrich those lists using over 50 different data sources. Sales teams, marketing teams, and even recruiting functions were using it for exactly that.
Then November 2022 happened. ChatGPT launched, and with it came the GPT-3.5 model. And that changed everything.
For the first time, there was a low-cost, easily integratable generative AI API that could be dropped into these workflows to produce high-quality outputs.
Clay, and tools like it, transformed almost overnight from list-building and enrichment platforms into full go-to-market workflow platforms.
Suddenly, you could build a list of contacts, enrich it with data from dozens of sources, run that data through an OpenAI, Claude, or Gemini model to generate personalized research or messaging, and then export the whole thing directly to your CRM or sequencer. End-to-end. Automated.
Two things made this resonate so strongly with the market.
The first was one-to-one personalization at scale. Before this, go-to-market teams basically had two options: send a small volume of highly personalized messages that required hours of manual research per contact, or send high-volume generic messaging and hope something landed.
The emergence of generative AI in these workflows made a third option possible: genuine personalization, at scale, without the manual research burden.
The second was the commoditization of go-to-market tactics. The half-life of any given sales play had been shrinking for years. What gave you a competitive edge for an entire quarter a few years ago was now lasting weeks, sometimes days.
The old playbooks were falling flat. Teams needed a way to spin up new campaigns and plays in days, not weeks, and that's exactly what this new generation of tooling made possible.
Dalip Jaggi
How organizations are restructuring around this role
If you look at how a typical go-to-market function was structured before this shift, you'd see something familiar: large SDR teams armed with ten or more tools for research, prospecting, and outreach, all working to book meetings for AEs, with revenue operations or marketing operations serving as the connective tissue holding everything together.
That structure is changing.
Today, we're seeing smaller groups of go-to-market engineers using a tighter stack, typically something like Clay, a CRM, and a sequencer, to run outbound at scale and feed a leaner sales team. The ratio of engineers to output has shifted dramatically.
And the reason is straightforward: in the old model, scaling revenue meant scaling headcount. With go-to-market engineers leveraging AI and automation, the ceiling on output is largely a function of creativity, not team size.
That's a meaningful shift. The only thing stopping you from scaling your go-to-market motion is how many good plays you can think of and execute. That's a very different constraint than "how many SDRs can we afford to hire."
The two organizational models we're seeing
Across the companies adopting go-to-market engineering, two primary models have emerged for where this function lives.
Model one: go-to-market engineers inside revenue operations
Companies like Anthropic, Intercom, and Canva have taken this approach. It works well because revenue operations already owns data pipelines and data quality.
Embedding a go-to-market engineer there removes the translation layer between identifying a problem and building a solution. There's no "here's what we need" followed by a lengthy back-and-forth before anything gets built. The person who understands the problem is also the person building the fix.
Model two: go-to-market engineers inside the growth organization
Companies like Verkada, Rippling, and Ramp have gone this route. The logic here is that embedding technical talent alongside demand generation unlocks ideas that traditional marketers might never attempt.
Programmatic landing pages, real-time intent harvesting, dynamic campaign logic: these kinds of outputs become feasible when you have someone who can actually build them sitting in the same room (or Slack channel) as the people thinking about the pipeline.
Both models work. The right fit depends on where the biggest bottlenecks are in your organization and where the go-to-market engineer can have the most direct impact.
Matteo Tittarelli
Who's becoming a go-to-market engineer?
The backgrounds are genuinely diverse, and that's one of the more interesting things about this role.
You'll find product designers with startup experience and a DIY mindset. You'll find rev ops professionals who already understand the go-to-market context and are learning to build on top of it. You'll find AI engineers who have a genuine interest in sales and marketing and want to apply their technical skills in a commercial context.
But across all of those backgrounds, a few common traits show up consistently in the best go-to-market engineers I've seen.
They think in systems. They don't approach problems in isolation. When they're solving a prospecting problem, they're thinking about how it connects to the enrichment layer, the messaging layer, the CRM, the sequencer, the feedback loop. Everything is part of a larger system.
They have a go-to-market context. Whether they've worked in a customer-facing role, in rev ops, or in marketing operations, they understand what go-to-market actually means in practice. They know what a good lead looks like. They know what makes a message land.
And they're creative. This might be the most important trait of all. The ability to bridge raw data and creative ideas, to look at a dataset and imagine a campaign that nobody else has run yet, that's what separates a good go-to-market engineer from a great one.
What's coming next
This is where things get genuinely interesting. These are my views, not certainties, but they're grounded in what I'm already seeing emerge.
Distribution is the new moat
AI has dramatically lowered the barrier to entry for building software. Engineers can ship faster, features can be replicated more easily, and the product advantages that used to take years to build are becoming easier to close the gap on.
As a result, I think we'll see more companies treat their go-to-market motion itself as their primary competitive differentiator. How you reach your market, with what data, through what channels, and with what level of precision, that becomes the edge.
This connects to what I'd call "go-to-market alpha": approaching segmentation and outreach using unique data points that AI can infer, rather than relying solely on static firmographic data.
Instead of filtering for "restaurants in a given zip code," you might look for restaurants that have recently launched on DoorDash, or that have a rating above a certain threshold, or that show some other behavioral signal that an AI model can surface. That kind of precision wasn't feasible before. It is now, and it's only going to become more accessible.
The shift toward agentic workflows
Right now, most go-to-market engineering work is deterministic: you build a workflow, it runs the same way every time, and you iterate on it manually based on what you observe.
I think we'll see a meaningful shift toward agentic workflows, where AI agents handle more of the actual execution within those workflows. But I don't think deterministic workflows disappear. What I expect to see is a hybrid model: agents embedded within deterministic frameworks, so you get the flexibility and intelligence of an agent with the guardrails and structure of a defined workflow.
A research agent might handle one stage of a workflow before handing off to a more rigid automated process, for example.
Reinforced learning in go-to-market workflows
Today, the feedback loop in any go-to-market workflow runs through the go-to-market engineer. They gather data on what's working, make adjustments, maybe run an A/B test, and iterate.
I think we'll see workflows that can capture feedback in real time and incorporate it into their own logic, essentially improving themselves continuously rather than waiting for a human to intervene. That's a significant shift in how these systems operate.
AI-driven segmentation
Rather than relying on firmographic data points that exist in databases, we'll see AI agents go out and find data points that don't exist anywhere yet. As the cost of AI continues to fall, the ability to generate unique, inferred signals about individual companies will become a core part of how go-to-market teams think about who to target and when.
Go-to-market engineers building rep interfaces
This one is a bit further out, but I think it's coming. Today, the data that reps need and the actions they need to take are spread across multiple tools and platforms.
The interfaces themselves are rigid and one-size-fits-all. I think go-to-market engineers will increasingly get involved in building the UI layer that sits in front of their back-end systems, using tools like Replit or Lovable to create custom interfaces that reflect how their specific team actually works.
The go-to-market engineer has the business context to build these interfaces efficiently, and the no-code and natural language-to-code tools now make it feasible without a dedicated front-end engineer.
Rebecca Stewart
The human element isn't going anywhere
One thing I want to be clear about, because I think it gets lost in conversations about AI and automation: the human element in go-to-market remains essential. Maybe more so than ever.
At Clay, we're actually anticipating that the rise of go-to-market engineering will lead to more emphasis on in-person events, not less. The automation handles the preparation and the follow-up. The human connection happens in the room.
Here's a concrete example. For our first user conference in San Francisco, we used Clay to take the attendee list, understand who was coming, whether they were current customers, and what their business objectives were.
That meant walking into the event knowing who I wanted to talk to and what mattered to them. After the event, we had a workflow where anyone you met could have their contact information photographed, sent to Clay, extracted, matched against Salesforce, and followed up with automatically if they weren't already in the system.
And the human-in-the-loop piece: I'd get a Slack message on Thursday morning saying, "You met this person at the conference, do you want to send them this email?" A simple yes or no. The system does the heavy lifting. The judgment call stays with a person.
That matters. CRMs will never capture everything. Business gets done over text messages that aren't logged, in hallway conversations, and over dinner. Without that context, even the best AI model can't tell you whether sending a follow-up email is going to strengthen a relationship or damage it. The human judgment layer isn't a limitation to be engineered away. It's a feature.
A genuinely new kind of role
Go-to-market engineering is the first AI-native career. The role was shaped by AI from the start, not adapted to accommodate it after the fact. And the people who thrive in it are the ones who can hold both sides of the equation: the technical ability to build systems and the creative, commercial instinct to know what those systems should actually do.
If you're thinking about where this fits in your organization, start by looking at where your biggest go-to-market bottlenecks are. Is it the quality of your targeting data? The speed at which you can launch new plays? The personalization of your outreach? The coherence of your follow-up across channels? Those are the problems a go-to-market engineer is built to solve.
The companies that figure this out early are already pulling ahead. And the gap between them and everyone else is only going to widen.
Want more on this topic?
AI-native RevOps runs through the San Francisco's summit agenda, from forecasting to stack architecture.
AI for GTM Summit is co-located with the RevOps Summit in San Francisco. Ashleigh Bilodeaux (GTM AI Engineer, Zoom) opens the main stage with "The rise of the GTM AI engineer."
Then the RevOps core: "Can AI predict revenue better than your sales team?" with Stephanie Ucko of Impact.com (3:45 pm), plus the panel on the architecture behind AI-native GTM.
San Francisco, September 22 and 23.
Revenue operations alliance insider
Thank you for subscribing
Level up your revenue operations alliance career & network with revenue operations alliance experts.
An email has been successfully sent to confirm your subscription.
Follow us on LinkedIn
