top of page
Search

"Expert in the Loop" Is the Marketing Advantage AI Can't Replace

  • Writer: Dave Hatch
    Dave Hatch
  • Jul 31
  • 6 min read

Updated: Aug 1

In what seems no time at all, AI has moved from novelty to infrastructure. Especially in the marketing profession.

 


Copy drafts, media plans, audience segments, creative variants, even full campaign briefs now get a first pass from a machine before a person ever touches them. Adoption has climbed fast: 97 percent of marketing leaders now use AI in their daily creative work, and 99 percent plan to increase AI investment this year, according to Canva's “State of Marketing and AI Report 2026” (https://www.canva.com/newsroom/news/marketing-ai-report-2026/), conducted with The Harris Poll. The question is no longer whether AI belongs in marketing. It's who stays in the loop, and why that still matters.


Our hypothesis is that the organizations getting the most value from AI aren't the ones that removed people from the process. They're the ones that moved the right people to the right point within the process. There's a meaningful difference between a "human in the loop," someone present to review or approve, and an "expert in the loop," someone whose qualifications, knowledge, experience and judgment improves the output. That distinction plays out differently depending on where you sit in the organization.


The C-Suite: Governance, Risk, and the Cost of Getting It Wrong


For CEOs, CFOs, and other C-suite leaders, AI in marketing has become a governance question. Every AI-generated claim, image, or customer interaction carries brand, legal, and reputational exposure the moment it leaves the building. A hallucinated statistic in a press release, a tone-deaf ad generated without cultural context, a pricing claim that isn't quite accurate: these are the failure risks that executives often miss until it is too late, or lose sleep over knowing they cannot be monitored, and they compound quickly at scale.


This is where the "expert in the loop" argument lands hardest at the top of the house. A generic review step, someone skimming for typos, doesn't catch a subtle factual error or a compliance issue. An expert reviewer, someone who understands the product, the regulatory environment, and the brand's history of near misses, does. Consumer sentiment backs this up: 31 percent of consumers say visible AI-generated marketing content makes them trust a brand less, while only 7 percent say it makes them trust a brand more, according to December 2025 research from [Klaviyo and Datalily] (https://www.klaviyo.com/solutions/ai/consumer-trust-in-ai). That's a governance problem before it's a creative one.


The C-suite's job, then, is to decide and enforce where autonomous execution has earned the right to run unsupervised and where it hasn't. That means building real checkpoints into workflows and staffing those checkpoints with people who have enough domain expertise to catch what the AI models missed. It also means treating AI governance as a board-level topic alongside cybersecurity and financial controls, because the risk profile is comparable even if the headlines haven't caught up yet.


Commercial Leadership: Speed Without Losing the Signal


Sales, marketing, and business development leaders live in a different tension: AI's promise of speed and scale versus the risk of producing more content and outreach that says less. A commercial leader's incentive is pipeline and revenue, and AI genuinely delivers there. Multi-agent systems can now research accounts, draft outreach sequences, adjust customer journeys in real time, and surface intent signals faster than any team could manually. That's a new operational reality in 2026.


But commercial leaders are also the ones who hear directly from the field when something feels off. For example, when a prospect receives three contradictory messages, or a campaign that technically hits its metrics but generates complaints instead of pipeline. The expert-in-the-loop model matters here because volume without judgment just produces noise faster. The data supports a hybrid approach: 73 percent of high-performing marketing teams combine AI with substantial human editing rather than relying on raw AI output, according to “Fueler's 2026 AI marketing statistics report” (https://fueler.io/blog/ai-marketing-statistics-updated), this is a pattern that holds because it protects conversion quality, not just output volume.


The practical implication for commercial leadership is to resist the temptation to measure AI success purely by throughput. Set targets for AI-assisted programs that include quality and trust metrics, not just volume:

  • Response rates

  • Prospect / Deal quality scoring

  • Customer sentiment scores

  • Conversion costs

  • Lead acceptance rates by sales


And make sure the experts who understand what a good customer conversation actually sounds like, senior account executives, veteran sales engineers, product marketers with deep buyer knowledge, are positioned to shape and correct AI output before it reaches a prospect.


Marketing Practitioners: From Producer to Editor and Strategist


For the people actually doing the work - copywriters, designers, demand gen managers, brand strategists - the shift is the most personal and, for many, the most disorienting. AI has taken over a meaningful share of first-draft production: headlines, ad variants, email sequences, even rough creative concepts. That can feel like a threat to the craft. It's more accurate to describe it as a “relocation of value”. The practitioner's job is moving from producing the first version to being the expert who makes the AI's version actually good. This poses a new paradigm for the way we think about entering the marketing profession. The traditional path of starting in some form of “marketing assistant” role is no longer viable. Marketers may start coming from different areas of the business, where learning about the business, the market, the selling conditions are all part of the toolset that is required before an “expert-in-the-loop” role can be achieved.


This requires a different skill set than pure execution. It requires taste, an ability to spot when something is technically correct but strategically wrong, and enough command of the brand's voice and audience to know exactly what to fix and why. Cognitute's “Agentic AI in Marketing guide for CMOs” (2026) (https://www.cognitute.org/insights/agentic-ai-in-marketing) frames the winning approach as preserving "the human creative and strategic intelligence that no agent can replace," even as autonomous agents take on more of the execution itself. Practitioners who lean into that role, becoming the discerning editor rather than resisting the tool or blindly accepting its output, are the ones whose value increases as AI capability and usage expands.


There's also a trust dimension practitioners are uniquely positioned to manage. Consumers have been fairly clear that they notice and often resent obviously AI-generated marketing: 78 percent say they would rather see ads made by people, even if AI could produce technically better ones, and 87 percent say the best advertising still requires a human touch, per the same Canva/Harris Poll report cited above. That preference isn't going away because the technology gets better. It means the practitioner's ability to inject genuine craft, specificity, and point of view into AI-assisted work is the thing that keeps the output from reading as generic.


The Common Thread


Across all three groups, the pattern is consistent. AI is extraordinarily good at generating options, drafts, and first passes at a speed and scale no human team can match. What it doesn't reliably do is know which option is right for your brand, customers, or your risk profile. That judgment comes from having the right expert positioned at the point in the process where their specific knowledge changes the outcome.


That's the real difference between human in the loop and expert in the loop. The former is a compliance posture. The latter is a competitive advantage. Organizations that treat AI oversight as a checkbox will get AI-scale output with human-scale problems. Organizations that deliberately place domain experts, executives who understand risk, commercial leaders who understand the customer, practitioners who understand craft, at the moments that matter will get something better: AI-scale output with human-scale judgment behind it.


For marketing leaders navigating this shift, the practical starting point is an honest audit of your own workflows: where does expertise currently sit in your AI-assisted processes, and is it positioned early enough, and with enough authority, to actually shape the outcome? That answer will look different for every organization. But asking it now, while the workflows are still being built, is far easier than retrofitting judgment into a system after it's already scaled without it.


David Hatch is Founder and Owner of CMO in Residence, where he advises companies on marketing leadership, strategy, and organizational design.


---Sources---


 
 
 

Comments

bottom of page