How we use AI at The Growth Syndicate.

Adoption is high. Capability is low. The gap is widening.

The pace of change is relentless, and nowhere more so than in B2B tech, where commercial and marketing roles are being redrawn by the month. Any company denying AI's impact is going to fall behind. And if you're hiring a team or an agency that can't actually wield it, you're either missing out or overpaying.

Our own research surfaced a harder truth. When we surveyed 110 B2B marketing leaders, belief in AI's potential was near-universal, yet far fewer could actually extract value from it.

Meanwhile 63% worry AI is increasing sameness and eroding differentiation, and skills, not budget, are the number-one barrier to progress. Most teams are stuck in what we call the dangerous middle: not creative or strategic enough to differentiate on taste, and not technical enough to extract real leverage from the tools. They buy tools before building the skills, systems, and governance to use them, solving the problem in exactly the wrong order.

We built our operating model to be the opposite of that.

Humans lead. AI executes.

Everything we do runs on a simple split: people own the thinking, and AI does the execution.

We like working in complex industries, and we go deep on the parts of the job a model can't reach. How a market actually moves. Why a customer buys, beyond what they'll admit in a survey. The politics inside a buying committee. And the judgment that only comes from solving the same kind of problem, many times over. That depth is what makes the work good, and it lives with people. So that's where we keep it.

We also don't underestimate what AI can do to the mechanics. It can accelerate and automate large parts of the GTM function, and we know how to set those systems up and run them across the whole operation: research and market intelligence, content generation, go-to-market systems and agentic workflows, the knowledge infrastructure underneath.

In the right hands, with expert direction on top, AI stops being a liability and becomes real capability. The combination gives us a level of speed and reach most in-house teams and generalist agencies can't match.

From AI-assisted to AI-restructured

There's a real difference between using AI and being changed by it. In our research, three maturity levels kept showing up, and the distance between them isn't linear. It compounds. We built the agency to run at the far end of it.

This is what we mean by AI-restructured. It's about how the work is actually organized. Concretely, that has meant:

• Re-architecting workflows instead of decorating them. For every major workflow, we know where AI runs the job by default, where it assists, and where a human leads.
• Scaling through infrastructure instead of headcount. We automate the low-judgment work and put that capacity into strategy and execution that moves metrics.
• Building systems, then handing over the keys. Documented and templated, so they keep running without us.

The six principles that guide our work

How this shows up in our work

Six areas where AI has changed how we operate, each rooted in a strategic understanding of the client's context.

AEO and GEO: built and measured

B2B discovery is shifting under everyone's feet. Buyers, especially in expansion personas like sourcing, legal, and supply-chain ops, increasingly start their research inside AI tools and answer engines instead of traditional search. In a zero-click environment, Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) become first-order disciplines that sit alongside SEO, not underneath it.

So we don't treat this as a side workstream off SEO. It changes content strategy, channel mix, and campaign architecture all at once. We test how brands surface inside AI-generated answers, we measure citation presence in AI-driven discovery, and we structure content so answer engines can actually read and quote it.

We're honest about the volatility. The rules are still being written, so we approach AI-driven search with real caution and don't pretend it's a sure thing. We've built this into live client programs and measured it in practice.

Where humans stay non-negotiable

The faster AI executes, the more governance matters. Scale without guardrails just produces mistakes, and sameness, more efficiently. Some things never move to a model, however capable it gets.

• Original strategy. AI recombines existing patterns; it can't reframe a problem or read an unmet market need from first principles.
• Brand voice and positioning. AI can mimic a style but not the source of what makes a brand distinctive.
• Regulatory and factual claims. Anything carrying legal or compliance weight gets human verification, always.
• High-stakes, customer-facing commitments. Trust and relationship capital can't be automated.
• Quality judgment. Knowing what "good" looks like is a pattern library built over years, not a prompt.

Two more run underneath all of it. Managing hallucination and drift: grounding outputs in controlled knowledge, as the Oracle does, is the first line of defense. We verify sources instead of trusting AI-supplied citations, we map every workflow by how much human oversight it needs, and we watch for drift so a system that worked last month doesn't quietly rot. And protecting data and IP: we build client-specific knowledge bases with clear boundaries, so agents reach only the sources we name, not everything. We anonymize and redact where required, keep audit trails, and honor NDAs to the letter. We don't prompt with "use 500 sources." We say exactly what a system can reference, which lifts output quality and protects data hygiene.

Building systems clients can own

A lot of AI work creates dependency: clever automations only the agency understands, and only the agency can run. We think that's the wrong model.

When we build infrastructure (agents, automations, knowledge systems), we build it to be handed over. We document how it works, template the inputs, define the supervision model (who checks what, and when), and instrument it so whoever runs it can see what's working and improve on it. Success is a system that keeps compounding value after we've stepped back, with no dependency on us. It's the same thing we do internally: make our own work obsolete by building the system, then lead the system.

Why this is hard to do in-house

AI in marketing is moving faster than any in-house team can track. The models and the tooling change every week, and so do the techniques for using them. Testing all of it, working out what actually holds up, then implementing and scaling the winners, is a full-time job on its own. Most in-house marketers already have one: targets to hit, campaigns to ship. They can't run a continuous R&D function on the side too. A few big teams have someone dedicated to it. Most don't.

The alternative most companies reach for is worse. They hand the AI work to people fluent in AI but not in marketing. The output looks impressive, all slick automations and demo-ready workflows, and then it doesn't survive contact with a real pipeline. There's a word for this now: workslop.

This advantage is structural. Staying on the frontier is part of the job, not a side project. We test new tools and workflows on our own business first, where the only thing at risk is our own time, and we prove them across clients, use cases, and industries before they ever reach yours. No single in-house team can replicate that breadth. We see what works in apparel and what works in fintech, what scales and what only demos well. You get the winners, already proven, and you don't pay for the dead ends.

The people behind the work

The hard part of AI in marketing was never the AI. It's pairing real AI fluency with real marketing and sales craft, and very few people have both. Most of the impressive-looking workflows out there were built by people who understand the tools but not the work. Ours are built by people who understand both.

We're deliberately headcount-light and AI-heavy. Senior strategists architect the systems and own the thinking. Technically-fluent operators build and run them. And specialist human craft, in content, creative, and design, comes in exactly where taste is the deciding factor. It maps to the split our own research found: the people who win with AI are either strategic leaders or technical orchestrators, rarely stuck in the middle. Ours are both.

Take Adrian, our Head of Content. He's fluent with AI tooling and Claude Code setups, comfortable wiring up the technical side himself, but he's a linguist and a content craftsman first. That combination is the whole point: the technical ability to make AI do real work, in the hands of someone whose actual expertise is language, story, and persuasion.

You can see what that produces in our Marketing for Manufacturing Report, a deeply researched, high-quality piece of work that could not have been turned around as fast without AI doing the heavy lifting underneath expert direction.

Who we do this with

We're at our best alongside companies that take AI seriously in their own products and go-to-market. We've worked closely with AI-first and AI-native businesses that build AI deep into the core of what they sell, rather than bolting it on as a feature:

That proximity keeps us sharp. And it's reflected in our published work: we created the first dedicated State of AI in B2B Marketing report, surveying 110 leaders and interviewing ten expert practitioners, founders, and operators. We generate the primary research that the rest of the field reads.

AI has to improve outcomes. Looking impressive on a slide is not enough. We use it where we can measure the impact, and we drop it where we can't.

This shift is only getting started. The tools and the models keep changing by the week, and so do the limits. Which is why the durable advantage is people, not any single tool or workflow. People who stay deliberately agnostic about tools, because the real skill is being technical enough to build almost anything and clear enough to know what's worth building in the first place. The tools will keep changing. That adaptability is the moat.

Marketing is in the middle of an AI shift that will reshape how commercial teams operate. We're not here to shout about it from the rooftops, and we're not here to fear it. We treat AI as what it is: an accelerator and a force multiplier for the things human marketing leaders do best. The aim is higher quality, at superhuman speed.