Why AI Content at Scale Means Nothing Without Quality Controls in Place

In this week’s episode of Voices of Search, we spoke with Holly Enneking, Vice President of Marketing at Markup AI, about why most marketing teams are treating AI adoption as the finish line when it’s really just the starting line—and what it takes to actually scale content without sacrificing quality, compliance, or trust.

Holly brings nearly two decades of experience across brand, digital, and content strategy, having led marketing at companies ranging from large enterprises to early-stage startups. She rejoined Markup’s CEO—someone she’d worked with twice before—because the problem the platform is solving is one she’d felt firsthand throughout her career.

Today, Holly broke down why the definition of “good” content is harder to pin down than most teams admit, how guardrails can replace the chaotic manual review process that’s slowing teams down, and why chasing every new AI platform is a strategy that will always leave marketers behind.

Key Takeaways From This Episode:

• 85% of marketers using AI for content creation are doing so without real quality controls in place
• AI will write what’s expected—the marketer’s job is to bring what’s unexpected
• Good content quality breaks into two distinct categories: black-and-white deterministic checks and subjective judgment calls, and teams need to handle each differently
• Guardrails don’t have to be sophisticated—a checklist in a Google Doc is a legitimate and valuable starting point
• Unsanctioned LLMs being used across teams is one of the most overlooked security and compliance risks in enterprise marketing today
• Being present in every AI platform is not a strategy—it’s a recipe for diluted, low-quality content across the board
• AI is a multiplier of the experience and judgment marketers already have, not a replacement for it

The Gap Between Generation and Publication

The core problem Markup AI is built to solve is what Holly called the gap between content generation and content publication. Teams are using AI to produce content faster than ever, but the review process sitting between “generated” and “live” is still largely manual—and manual doesn’t scale.

“AI has made it so easy to produce content at scale,” Holly said, “but it hasn’t made that content high-quality or trustworthy.”

What fills that gap at most companies today is a patchwork of individual reviewers, each bringing their own standards, thresholds, and blind spots. The result is inconsistency at best and compliance risk at worst. Holly was clear that this isn’t just a productivity problem—it’s a trust problem, and for B2B companies especially, trust is the foundation of any impact on pipeline and revenue.

Markup’s answer is content guardian agents that live inside the workflows teams already use, scanning and scoring content against brand voice, preferred terminology, accuracy standards, and compliance requirements before anything goes live. 

The goal isn’t to replace the human reviewer—it’s to handle the deterministic, black-and-white checks automatically so that humans can focus their attention on the harder judgment calls.

Defining What “Good” Actually Means

One of the most significant moments in the conversation came when Holly addressed how few teams have actually defined what publish-ready looks like. “Good” as a content standard is, in her words, largely vibes-based—and that’s a problem when you’re trying to scale.

Holly drew a clear distinction between two types of content quality:

• Deterministic checks: Is the brand name written correctly? Are product names accurate? Are claims unsupported? Is the source credible? These have clear yes or no answers and should be automated
• Subjective judgment calls: Is this saying something unique in the market? If you swapped in a competitor’s name, would it sound exactly the same? Is there a real point of view here? These require human expertise and can’t be automated

The problem most teams run into is spending all of their manual review time on the first category—burning out before they ever get to the second. “Marketers are spending a lot of that manual time on the black and white,” Holly said, “and then they’re burned out by the time they get to the tough questions.”

The fix isn’t to hire more reviewers. It’s to automate the deterministic layer so that human attention can go where it actually matters.

What Guardrails Look Like in Practice

Holly was refreshingly direct about the fact that guardrails don’t have to be complicated to be effective. The concept can sound intimidating, but the starting point is simple: write down what has to be true every single time before content goes live, get everyone aligned on that list, and make it part of the workflow.

A practical guardrail framework might include:

• A documented definition of what “publish-ready” means for your team
• A checklist of non-negotiable brand and accuracy standards that apply to every piece
• Clear ownership of who has final say on content quality—Holly’s view is that it should be the marketer, not a committee
• Agreed-upon standards for brand voice that go beyond “we want to sound friendly” and get specific about phrasing, tone, and structure

“Setting the baseline, getting people’s involvement in that, having alignment around it, and then sticking to it—that’s the biggest piece,” Holly said. “Just making that a part of your workflow.”

She also flagged something she hears often: marketers who have generated dozens of content pieces that are just sitting in a folder because the review process feels too overwhelming to start. A clear guardrail framework doesn’t just improve quality—it removes the paralysis that stops content from ever crossing the finish line.

The Overlooked Risk Nobody Is Talking About

When asked about the most overlooked risk in AI-generated content, Holly didn’t point to hallucinations or brand inconsistency. She pointed to unsanctioned LLM use across teams.

When employees are pulling in their own AI tools—models and platforms that the business hasn’t vetted, approved, or even identified—the business has no visibility into what data is being fed into those systems. Customer data. Prospect data. Internal strategy documents. All of it potentially flowing into third-party models with no oversight.

Her advice for marketers and leaders:

• Know what tools your team is actually using, not just what they’re supposed to be using
• Set a clear policy around which platforms are approved and what data can be entered into them
• If your business has an AI mandate, pair it with explicit guidance on tools and data handling
• If you’re an individual contributor without that clarity, ask for it—the liability risk of getting it wrong is real

“If you’re a person at a company who doesn’t have that clarity, you should be asking those questions,” Holly said, “because the last thing you want to do is be the one that causes some sort of data problem down the road.”

AI Writes What’s Expected—Your Job Is to Write What Isn’t

One of the most important strategic points Holly made was about what AI is fundamentally incapable of producing on its own: a genuine point of view.

AI-generated content reflects what already exists. It synthesizes, aggregates, and mirrors the internet back at you in a coherent format. What it cannot do is say something unexpected—bring a perspective that is specific to your brand, your team’s experience, or your audience’s actual situation.

That gap is where marketers earn their keep. “AI is going to say what’s expected,” Holly said. “You need to say what’s unexpected, and that’s the piece that the marketer can bring to it.”

This matters more now than before AI because the bar for content cited in AI search results is rising. Schema and formatting still matter, but they’re table stakes. What actually gets surfaced is content that adds genuine value—content that answers a question in a way no one else could have.

Chasing Hype Cycles Is a Quality Problem

The conversation took a bit of a turn when Jordan raised something that most content teams have lived through: the constant pressure to shift strategy toward whatever platform or channel is having a moment. Reddit citations dominate AI results one week, LinkedIn authority the next, and Wikipedia the week after that.

Holly’s response was direct: teams need to be ruthless about focus:

• 80 to 90% of a team’s time should stay locked on the channels and strategies they’ve already committed to
• The remaining capacity can absorb genuine experiments—but not at the expense of core execution
• A strategy that’s constantly pivoting is a strategy that never generates the data needed to know what’s actually working

“Every marketing team is being asked to do more with less,” Holly said, “and teams that are getting their time divided by all of these spur-of-the-moment things are really going to struggle.”

The same logic applies to the proliferation of AI platforms. With somewhere between 20 and 30 prominent LLM models now in the market, the instinct to have a content strategy for every one of them is understandable—and completely unsustainable. Holly’s view is that the right question isn’t which platforms you should be in. It’s where your audience actually is, and whether you’ve earned enough presence there to know if it’s working before chasing the next thing.

The Fundamentals Are Still the Multiplier

Holly closed with something that runs through everything she’s built in her career: AI doesn’t replace marketing fundamentals. It amplifies them.

The marketers who get the most out of AI are the ones who bring a clear point of view on their audience, a strong sense of what quality looks like, and the judgment that comes from years of working across different types of businesses and teams. Feed those inputs into AI, and the output gets dramatically better. Skip them, and you’re just generating noise faster.

“AI is really a multiplier,” Holly said. “It’s taking your brain times 100, which is incredible. But good prompts in are going to get good content out. Bad prompts are going to get bad outputs out.”

The teams that will win the content quality challenge aren’t the ones producing the most. They’re the ones who defined what good looks like, automated the checks that don’t require human judgment, and freed up their best thinking for the work that actually differentiates their brand.

Voices of Search is a daily SEO and content marketing podcast hosted by Jordan Keone and Tyson Stockton. The show delivers actionable strategies and data-driven insights to help marketers navigate the ever-evolving world of search engine optimization and content marketing. New episodes air weekly, covering everything from technical SEO to AI discovery, featuring industry leaders and practitioners sharing real-world frameworks and proven tactics.

Subscribe to Voices of Search on Apple Podcasts, Spotify, or your favorite podcast platform. Follow Previsible on LinkedIn for updates and subscribe to the VOS YouTube channel for video episodes and clips. You can also visit the official VOS site to explore the full episode archive and submit your SEO questions for future episodes.