Over the past month, I’ve spoken with multiple clients looking for help articulating to their executive teams how they’re integrating AI into marketing workflows. B2B marketing leaders are in need of more effective ways to showcase AI impact to other organizational leaders and stakeholders, whether that’s through driving new efficiencies or increasing results from every dollar spent.
The question is starting to show up in a more direct form: “Can AI replace agency work?”
It’s a fair question. It reflects the real pressures AI is introducing around cost, speed and competitive advantage. As I’ve thought more about these recent conversations, my advice is to answer this question by focusing on how AI is changing the quality and impact of agency work, not as a replacement mechanism.
Because AI is a multiplier. The impact — good or bad — depends on who’s using it.
Key Takeaways
- AI is helping high performers work faster, sharper and more effectively. Lower-skill AI users also move faster, but often in the wrong direction.
- Adding AI alone won’t make a mediocre marketing system great. It will amplify the system that already exists, making its strengths more productive and its weaknesses more pronounced. True domain expertise is critical in AI-driven environments.
- In practice (e.g., SEO), AI produces outputs. Activities like strategy, data quality and interpretation still determine results.
The AI Amplifier Effect
If we’ve learned anything in the first two years of AI adoption, it’s that the technology is widening gaps between winners and losers.
AI increases output, but it doesn’t improve judgment. This dynamic is creating a noticeable divide between teams today.
In software development, for example, recent research from DORA on AI-assisted workflows shows a consistent pattern: Higher-skilled developers become significantly more productive, while lower-skilled developers often produce worse outcomes. Both are completing more work thanks to AI, but for developers lacking the judgment skills to identify when AI is leading them off track, that effort is headed in the wrong direction.
The same pattern is emerging in marketing.
High-performing marketing teams use AI to work faster, sharper and more effectively. These teams have clear strategy and positioning, clean and structured data, tight feedback loops and operators who know what good looks like. Organizations introducing AI in this type of environment accelerate everything that already works. Testing velocity increases. Insights compound. Execution scales more efficiently.
But the inverse is also true. Adding AI into workflows won’t make a mediocre marketing system great; it will amplify the system that already exists. Teams with vague positioning, fragmented data and inconsistent prioritization don’t suddenly become more strategic with AI. Instead, they produce more of the same issues at greater speed.
The result is marketing activity that looks productive but fails to drive outcomes.
Human Expertise Multiplies Marketing Impact: An SEO Use Case
This AI pattern is playing out across virtually all marketing disciplines, from paid media to PR to content and analytics.
SEO provides a clear example of the AI amplifier effect in action. AI accelerates many parts of SEO practices, but performance depends on how teams — including agency partners — direct, interpret and apply outputs.
- Strategy still drives performance. You can’t rely on AI alone to produce an effective SEO strategy. Across models, AI tends to generate similar strategic recommendations regardless of the specific business situation. Research from Harvard Business Review found that leading AI systems consistently give the same advice on strategic trade-offs, often aligning with familiar managerial trends and buzzwords rather than situation-specific logic. That means an AI output can sound highly tailored while quietly steering you toward a narrow set of generic approaches others are already using. A successful SEO or GEO strategy requires something different. AI can generate keyword lists and content ideas, but it doesn’t define where a brand can win, which segments matter most or how SEO connects to broader business goals.
- Data advantages matter more than ever. AI outputs reflect the data you give them. In practice, automated keyword research through tools like Google AI Max often surfaces irrelevant or low-value opportunities without the context to filter and prioritize them. Even sophisticated tools get keyword research wrong, all of the time. Strong SEO programs rely on proprietary data sets, paid tools and historical performance insight — alongside SEO trained humans to separate signal from noise.
- Interpretation is the real differentiator. AI can produce audits and analyses, but it can’t reliably interpret a full website, competitive landscape or performance history. Effective SEO requires synthesizing inputs from multiple tools and translating them into clear, actionable direction.
- Technical SEO requires prioritization. AI can surface hundreds or even thousands of potential issues. The challenge is deciding what matters most and what to tackle first within budget. What will move rankings most? Should a page even be crawled and indexed? What should the canonical be? Technical SEOs specialize in prioritizing needs based on non-cookie-cutter business strategies, with an understanding of the website stack and implementation effort.
- Content optimization demands nuance. AI applies generalized best practices. Performance depends on how SEO content adapts those practices to specific industries, audiences and competitive environments. Experience determines where to push, where to adjust and where to take a different approach entirely.
- Report on outcomes, not outputs. AI produces a reporting framework, but it doesn’t provide insights or recommendations. SEO results come from prioritization, alignment and execution against clear goals.
The AI amplifier effect isn’t unique to SEO. In paid media, AI can now optimize bids and generate targeting recommendations, but performance still depends on strategy, budget allocation and interpretation of results. In PR, AI can help draft messaging and identify media trends, but it can’t replace narrative judgment or reporter relationships. In creative and content, AI can scale production, but it can’t define what will resonate or differentiate a brand in market.
The same results show up everywhere in B2B marketing: AI accelerates execution, but human expertise and judgment determine impact.
The Risk of Getting AI in Marketing Wrong
Without context or expertise, AI can misdiagnose problems, create unnecessary urgency and send teams in the wrong direction.
Walker Sands sees three risks showing up consistently:
- False confidence. AI outputs are polished and coherent, even when they’re wrong. That makes it harder to spot issues before they go live or scale.
- Cognitive debt. When teams rely on AI to handle upfront thinking, they defer deeper analysis. Over time, that leads to weaker strategy, thinner customer understanding and greater dependence on AI-generated recommendations.
- Content bloat. Teams generate more assets than they can realistically evaluate or improve.
Our team saw these types of AI challenges play out with a client just the other day. The client, a $300M+ legal tech company, raised concerns after an AI tool failed to process a page on its website. They logically assumed AI systems couldn’t access key web page content due to a large HTML-structured navigation menu at the top of the page.
In reality, though, the issue was a limitation of the question being asked in the AI tool and the token limit of the tool itself, not the site. And it was human involvement that understood and solved the problem.
Gaps like this highlight how easily AI can mislead even the most seasoned marketers. Without the expertise to interpret the output, our client might have made unnecessary changes, redirected resources and introduced additional problems.
This is the AI amplifier effect in action. The same AI tool produces dramatically different outcomes depending on who is using it and how they interpret its work.
The New Agency Model
Business leaders are right to sense that something is changing. But what’s changing is not the need for agency support. It’s how to evaluate strategy and results.
In AI-driven environments, the best agencies are defined less by what they produce and more by how they strategize: orchestrating tools, aligning data and applying judgment so AI outputs translate into real outcomes.
So when the question comes up — “Can AI replace agency work?” — the answer is straightforward: AI can replace pieces of execution, but it can’t replace the systems and humans that turn that execution into results. Those systems and humans are exactly where the right agency creates value.
In this way, AI is not lowering the bar. It’s raising the ceiling. AI doesn’t replace marketing expertise. Rather, it makes its presence (or absence) impossible to ignore.
Want more perspectives like this? Subscribe to the Walker Sands newsletter for practical insights on how marketing leaders are navigating AI, driving performance and building systems that deliver results.
