Why AI-generated content is only as good as the expert behind it — and what that means for growth-stage companies.
Founder & President, JTS Marketing
The emergence of large language models has disrupted marketing as both an industry and a discipline. Perhaps most noticeably through their ability to generate rich content at scale, LLMs have put a legitimate question on the table: what does the future of marketing look like — for internal teams and for the agencies that support them?
For the organizations we work with — established, growth-stage companies, typically in the 500-to-1,000-employee range, sometimes earlier — that question has direct consequences. Unlike the biggest players, growth-stage companies cannot staff for every experiment or absorb every wrong turn. They need to be judicious. They need outside judgment they can trust. And right now, many of them are being sold a version of AI adoption that skips the most important part.
Our view: AI can meaningfully accelerate and scale content development. What it cannot do — not without an experienced expert directing it — is replace the judgment, contextual understanding, and category fluency that great marketing actually requires.
Marketing is, and will always be, about relevance. It’s about capturing someone’s attention, keeping it, and getting them to know, believe, and ultimately do something. LLMs have become extraordinarily good at producing content that resembles that work. But, there is no substitute for the understanding that comes from years of category immersion — or for an understanding of emotional stakes, which remains the key ingredient in making a prospect care about what you’re saying.
Yes, our professional lives and choices are filled with emotion. You can prompt an LLM to be empathetic. You cannot prompt it to have lived in a category.
More practically: LLMs are only as reliable as the expertise guiding them. Without a knowledgeable hand on the wheel, the risk of error in technical B2B content is real and consistent. We have seen it firsthand. During a recent effort to develop a media pitch in the credit and lending space, we asked an LLM to identify and cite relevant articles using a specific set of criteria. It produced a fully fabricated New York Times article — complete with a plausible headline and a URL that led nowhere. On another occasion, we used an LLM to help shape a thought leadership piece built around an actual CEO interview with supporting sources. The model synthesized across those sources and generated a quote, attributed to the executive, that he had never said nor would have said.
In both cases, the error was caught because someone with editorial judgment and category knowledge was in the room. Without that, either piece goes out — and takes the company’s credibility with it.
This is not an argument against AI. We use it every day and consider it essential for many things. It is an argument for understanding precisely what it requires to work well: expert direction, editorial oversight, and deep enough domain knowledge to recognize when the output is wrong.
AI is not end-to-end. It’s middle-to-middle. It compresses the distance between brief and draft. It does not replace the judgment about what matters, what to lead with, or why a reader should care.
There is a structural reality about in-house marketing that rarely gets discussed plainly. Most career in-house marketers are hired to manage — internal stakeholders, staff, agency relationships, approval workflows, vendor contracts. They are generalists by organizational necessity, and frequently they are not practicing the crafts themselves: writing, media relations, content strategy, editorial development. This is not a criticism; it reflects how organizations are structured. But, it matters when evaluating what an in-house team can realistically deliver on its own.
The agency model works differently. A career agency practitioner, over a comparable tenure, gets immersed in many more companies than their in-house counterpart — often across the same category or adjacent ones. They are not managing a process from a distance; they are “in the studio” doing the work, repeatedly, across different organizations, different founders, different GTM contexts. The pattern recognition that accumulates from that exposure is not something any single organization can develop internally, regardless of how talented its team is.
The combination of in-house and agency in the right measure and moments delivers an exponentially better product and value.
At JTS, we have worked with dozens of fintech clients and their partners in payments alone — one of the most nuanced and specialized areas of financial services. That breadth is precisely what allows us to bring immediate, calibrated judgment to a new engagement rather than spending the first six months learning the category.
The buyer doesn’t know enough about a space to know what questions to ask. They use an AI model to build a mental model of the category before they engage with any vendor content directly.
Once the buyer understands the category, they ask the model to help them identify the leading players — and the key differentiators between them. The model’s answer shapes the initial shortlist.
Later in the process, buyers use AI tools to synthesize large volumes of vendor content, analyst reports, and third-party commentary into a comparative view.
At every stage, the AI model is drawing on a corpus of content it has been trained on, retrieved in real-time, or both. The companies whose thinking, expertise, and perspective is most thoroughly represented in that corpus have a structural advantage — regardless of how their website ranks on traditional search.
The companies whose thinking is most thoroughly represented in the AI corpus have a structural advantage. This is not the same thing as ranking well on Google — and conflating the two is costing B2B marketers visibility they don’t know they’re losing.
The question of whether to build in-house or work with an outside expert is not binary — and treating it as such is one of the more expensive mistakes a growth-stage company can make. Increasingly, we see smart, early-stage and pivoting companies avoiding this pitfall and leaping ahead by hiring fractional CMOs that are steeped in their category.
Building internal teams around recurring, predictable marketing needs makes complete sense. The daily cadence every organization requires is well-suited to in-house execution. But in a quickly accelerating business environment where everyone is using AI to increase the volume of content, no growth-stage company can “shift-hire” its way through all of its needs — certainly not its most time-sensitive ones.
The economics are instructive. A $150,000 marketing hire, fully costed (TCO) across benefits, payroll taxes, overhead, and management time, typically runs $225,000–$277,000 annually — a multiplier of 1.5 to 1.85 — before accounting for hiring or attrition costs. The equivalent investment in agency fees provides access to three to five specialized practitioners on a fractional basis, immediately available, managed by senior leadership with direct accountability for results.
The time dimension – TTV – matters just as much. The insourced path to full productivity — hiring, onboarding, ramp-up to genuine category fluency — typically runs five to nine months. The equivalent agency engagement is operational in two weeks – six for a full GTM lift.
Bottom line: If you tried to hire for every strategic initiative, you would miss your window before the first offer letter cleared legal.
The organizations that navigate this well build lean internal teams around their recurring needs, use AI to accelerate execution, and bring in external partners with genuine category depth for the work that actually moves the needle.
That is the model we work from at JTS. Not as vendors. As partners.
Because turning up the volume isn’t the same as turning out “hits”.
25+ years leading B2B marcomms for Mastercard, FedEx, and Nokia. PRWeek 40 Under 40. Juan advises investors and leadership teams on marketing strategy, GTM alignment, and the emerging discipline of Generative Engine Optimization for B2B technology companies.
Thinking about your GEO strategy? JTS Advisory helps B2B fintech companies build the content infrastructure that earns AI citation.
