A Manifesto on The Human-AI Sweet Spot, For Now Anyway…
In late 2022, I watched ChatGPT go viral and felt something I hadn’t experienced in years: genuine uncertainty about the future of my craft.
I’ve been a storyteller my entire career. Words are my tools. Narrative is my trade. And suddenly, here was a machine that could string sentences together faster than I could type my morning coffee order. The marketing blogs were already declaring the death of copywriters, designers, and anyone else whose work could theoretically be automated.
So I did what any reasonable person would do when facing an existential professional threat: I dove headfirst into it.
Over the past three years, I’ve logged somewhere north of 8,000 hours working with generative AI tools. LLMs like ChatGPT, Claude, and Gemini. Image generators like Midjourney, DALL-E, and Stable Diffusion. Video tools that have evolved from laughable to genuinely impressive. I’ve pushed these systems to their limits, found their edges, broken them in creative ways, and slowly developed a working philosophy about where they fit in the storytelling toolkit.
What I’ve learned has fundamentally shaped how we approach work at Vnaqlr and it’s probably not what you’d expect from someone who’s spent that much time with the technology.
The Hammer Problem
Here’s the metaphor that keeps coming back to me: AI tools, as they exist today, are hammers.
Really good hammers. Impressive hammers. Hammers that would have seemed like magic five years ago.
But they’re still hammers.
And if you’ve ever tried to cut glass with a hammer, you know how that ends. It’s not that the hammer is useless—it’s exceptional at what it’s designed for. But precision work? Delicate work? The kind of work that requires nuance and a steady hand, they are unequivocally not. So what happens when you use a hammer on something that requires subtly? You end up with a broken product, because that’s not what hammers do.
Let’s dig in a little to the limitations and the processes.
When I use an LLM to clarify my thinking, most often I am looking to have a conversation that helps me untangle a half-formed idea. When I use it like this, it’s remarkable. The tool excels at this. It’s patient, it’s responsive, and it helps me think out loud in a way that’s genuinely useful. Same with using Midjourney or any of the other image generation tools, to explore a visual direction, to generate twenty rough concepts in an hour so I can identify what’s resonating before I bring in a human designer for the precision work.
These are hammer tasks. And the hammer crushes them.
But when I need to write something that will make a reader feel something? When I need to capture a brand voice so specific that only three people in the world would recognize it as authentic? When I need to make a creative choice that breaks a rule because breaking that particular rule serves the story, I have to pivot.
That’s glass-cutting work. And every time I’ve tried to use the hammer for it, I’ve ended up with shards on the floor.
What 8,000 Hours Actually Revealed
Let me be specific about what I’ve learned, because vague AI takes aren’t useful to anyone.
Where AI genuinely excels:
First drafts of structured content. If I need a framework; an outline, a list of considerations, a rough structure for something I’m going to rewrite anyway, AI is remarkably efficient. It gets me 60% of the way there in 5% of the time. That is something that gives me more time to spend on the delicate work, the work that really pays out for the client.
Research synthesis. I can feed an AI tool a mountain of information and ask it to identify patterns, summarize key points, or highlight contradictions. It won’t catch everything a human expert would, but it dramatically accelerates the early stages of understanding a new topic. Last month, I used Claude to synthesize forty pages of competitor messaging before a brand positioning project. What would have taken me a full day took two hours, and I walked into the strategy session with clearer patterns identified than I would have found on my own.
Ideation volume. When I need quantity, like fifty headline options, twenty angle variations, a hundred different ways to approach a concept, AI can generate that volume in a minute. Most of it will be mediocre or derivative, but buried in that pile are usually two or three sparks worth developing. I think of this as panning for gold. The AI gives me a riverbed full of sediment. My job is to know gold when I see it.
Editing and refinement prompts. Asking AI to identify weak sentences, flag passive voice, or suggest tighter phrasing can be genuinely helpful. Not because its suggestions are always good, but because they force me to defend my choices or recognize where I was being lazy.
Where AI consistently falls short:
Original creative vision. This is the big one. A recent Wharton study put numbers to something I’d felt intuitively: when groups use AI for creative ideation, only 6% of the resulting ideas were unique. In the human-only group? Close to 100% “unique” ideas. AI is drawing from the same well every time, which means it’s constantly pushing toward the center of what’s already been done.
This study is not taking into account that even humans are consistently drawing on the same human well, and reshuffling old ideas and calling them new, so I take it with a grain of salt, but what it is telling us to be aware of is the unconscious desire to profess something as unique. There is an old adage, there is nothing new under the sun. Everything we create is a mish mash of something that came before, inspired by the things we have seen and heard. What is unique is how we tell that story, our own filigree that we add to the veneer. AI doesn’t do this well, and thus the human remains objectively its better.
Emotional authenticity. AI can mimic the structure of emotional writing—it knows where the beat should land, where the vulnerability should appear. But there’s a hollowness to it that readers detect, even when they can’t articulate why. The words are right; the soul is missing. I’ve run experiments where I show clients two versions of the same piece—one AI-generated and polished, one human-written. They can’t always explain the difference, but they almost always prefer the human version. There’s something in the rhythm, the unexpected word choice, the tiny imperfections that signal a real person was here. AI is PERFECT, every time, from the punctuation to the word choice, where the human falters, the human reader can tell, nearly every time.
Brand voice precision. I’ve spent hours trying to train AI models on specific brand voices. They get close. They get the general vibe. But they miss the micro-choices; the specific words a brand would never use, the rhythm that makes the voice feel like a person rather than a persona. One of our cannabis clients has a voice that’s irreverent but never crude, knowledgeable but never preachy. AI can hit one note or the other. It can’t walk that tightrope. So even if I used AI to get 75% of the way there, it still requires the careful crafting and unweaving of the mess that AI attempted.
Strategic risk-taking. The most memorable marketing breaks rules. It zigs when the category zags. AI is trained on what already exists, which means it’s structurally incapable of generating the truly unexpected. It can remix. It can’t revolt. It can recognize the pattern, but it will almost never BREAK it. That is a human agency that AI will probably struggle to reach for sometime.
The Sea of Sameness
Here’s what keeps me up at night—and what I think every marketer should be paying attention to as we move into 2026.
AI-generated content is flooding every channel. Blog posts, social captions, email sequences, ad copy; the sheer volume of AI-assisted content has exploded. And because these tools are all drawing from similar training data, using similar optimization patterns, and being prompted by people who read the same “how to use AI for marketing” articles…
Everything is starting to sound the same.
I call it the sea of sameness. Scroll through LinkedIn for five minutes and you’ll see it. The same cadence. The same structures. The same “I used to think X, but then I learned Y” frameworks that AI loves to generate because they perform well algorithmically. The same emoji patterns. The same “Here’s what nobody tells you about [topic]” hooks.
It works, technically. The content checks all the boxes. It’s optimized, it’s clear, it’s professional. It works and AI uses it because that’s how people were being seen on YT, on Facebook, and Instagram before Chat GPT and the LLM cousins came roaring to the fore. So its based on human understanding of psychological marketing, but its lacking a certain je ne sais quoi.
It’s also completely forgettable. You read it and your brain almost never remembers it. In one ear and out the other. Curious isn’t it? It’s as though our brain has a filter for the false.
I ran an informal experiment last month. I pulled twenty “thought leadership” posts from my LinkedIn feed and asked three colleagues to guess which were AI-assisted. They correctly identified eighteen of them. Not because the writing was bad—it wasn’t. It was competent, structured, polished. But it had a certain… sameness. A predictability in the rhythm. An absence of the weird, unexpected turns that make human writing memorable. Even the most polished pieces with personal reflections landed weirdly, as if the body itself was rejecting it like an immune system response.
What intrigues me is how we can stop blindly using the tools, and refine our usage, while also dramatically reducing our usage. Pumping out volume is killing the vibe, and more than likely our water systems and much more. So how can we use it more responsibly and maintain our human-ness?
This is the opportunity hiding inside the AI revolution. As everyone races toward efficiency and volume, the brands that invest in genuine human creativity—the weird ideas, the risky choices, the authentic voices that can’t be replicated by prompting a model—those brands are going to stand out like beacons in a fog.
The bar for “good enough” content has dropped to the floor. Which means the bar for remarkable content has never been more valuable to clear.
How We’ve Found the Sweet Spot at Vnaqlr
I’ve spent three years figuring out where AI fits in our process, and here’s where we’ve landed:
AI as prep cook, human as chef.
In a professional kitchen, prep cooks handle the foundational work; chopping vegetables, measuring ingredients, organizing the mise en place. The chef takes those prepared elements and transforms them into something with intention, artistry, and a point of view.
That’s our model. AI handles the prep: research synthesis, first-draft structures, ideation volume, competitive analysis, data processing. Humans handle the cooking: creative direction, voice development, strategic choices, emotional resonance, the final 40% that separates forgettable from unforgettable. Most clients are not paying you for the grunt work, they couldn’t care less about that element. They care about the deliverable and most of that is crafted by the 40% aforementioned, so why wouldn’t you look to reduce the time spent on the elements the client doesn’t really and truly pay you for?
This isn’t about being pro-AI or pretending/hoping that AI doesn’t exist. It’s about being honest about what produces results. Our clients don’t hire us for efficient content. They hire us for content that moves their audience. That requires human judgment at the center.
We never ship AI output directly.
This is a hard rule. Every piece of content that leaves our shop has been substantially transformed by human hands. Not edited, transformed. If we can’t point to specific creative choices that a human made, choices that required judgment and taste and an understanding of the specific audience we’re serving, it doesn’t go out.
I’ve had team members push back on this. “But the AI output is good enough,” they’ll say. And sometimes it is good enough—technically competent, structurally sound, optimized for whatever metric we’re chasing. But “good enough” isn’t what builds brands. “Good enough” doesn’t create the emotional resonance that turns customers into advocates. “Good enough” is what everyone else is doing, which makes it the fastest path to irrelevance.
We use AI to accelerate, not replace, human thinking.
The best use I’ve found for these tools is as a thinking partner. Not to generate final content, but to help me clarify what I actually want to say. The conversation with AI often reveals gaps in my own thinking, assumptions I hadn’t examined, angles I hadn’t considered. Then I close the chat window and write the thing myself.
Michael and I talk about this constantly. The tools have made us better thinkers by forcing us to articulate our ideas more clearly. But the ideas themselves? Those still come from lived experience, from understanding our clients’ businesses, from years of pattern recognition in this industry. No AI has that.
We’re honest with clients about what’s AI-assisted and what isn’t.
This matters more than most agencies want to admit. When a client hires Vnaqlr, they’re hiring our perspective, our taste, our accumulated experience in their industry. If we were just prompting AI and polishing the output, they could do that themselves for a fraction of the cost. The value we provide is the human judgment layer, and we’re explicit about that.
What I’m Watching for in 2026
The tools are going to keep improving. That’s a given. But here’s what I think actually matters for marketers in the coming year:
The authenticity premium is going to increase.
As AI content becomes more prevalent and more detectable (both by audiences and by algorithms), genuinely human content is going to command a premium. Brands that can demonstrate authentic human creativity—not just claim it—will have a significant competitive advantage.
Strategic thinking becomes the differentiator.
AI can execute tactics. It cannot develop strategy. It cannot understand the nuanced position a brand occupies in a customer’s mind, or make the counterintuitive choice that breaks through category noise. The marketers who thrive will be the ones who can think strategically and use AI to accelerate execution of that strategy.
Voice becomes everything.
In a sea of sameness, distinctive voice is the ultimate differentiator. The brands that invest in developing a genuine, recognizable, impossible-to-replicate voice will own their categories. This is fundamentally human work—and it’s going to be more valuable than ever.
The hybrid model wins.
Pure AI content will underperform. Pure human content will be too slow and expensive at scale. The winners will be teams that figure out the right hybrid approach—using AI to handle what AI does well, reserving human creativity for what humans do irreplaceably.
How You Can Apply This
If you’re figuring out your own AI strategy, whether you’re a one-person shop or a marketing team of fifty, here’s where I’d start:
Audit your current workflow for hammer tasks vs. glass-cutting tasks. Be honest about which parts of your process are repetitive, structural, and scalable (hammer territory) versus which require judgment, creativity, and nuance (glass-cutting territory). Apply AI aggressively to the first category. Protect the second.
Develop a “minimum human touch” standard. What level of human creative involvement does a piece of content need before it represents your brand? Define this clearly. Make it a policy. Don’t let efficiency pressure erode it. This is where you slip into dangerous territory…
Invest in voice development. If you don’t have a documented, detailed brand voice—one specific enough that you could identify whether a piece of content matches it—you’re vulnerable to the sea of sameness. This is the strategic work that pays dividends for years.
Use AI as a sparring partner, not a ghostwriter. The most valuable thing these tools do isn’t generate content—it’s help you think more clearly. Have conversations with them. Argue with them. Let them challenge your assumptions. Then go create something yourself.
Stay skeptical of “AI-first” marketing advice. There’s a lot of content out there (much of it AI-generated, ironically) claiming that AI can do everything. It can’t. The people selling you on fully automated content creation have a vested interest in that narrative. Your competitive advantage lives in the stuff that can’t be automated.
Lastly, use them and use them often to determine how they fit for you and your business.
The Real Sweet Spot
After 8,000 hours, here’s what I’ve concluded: the human-AI sweet spot isn’t a destination.
It’s a practice.
It’s the ongoing discipline of asking, for every task: Is this a hammer job or a glass-cutting job? It’s the willingness to use AI aggressively where it excels while fiercely protecting the creative work that makes your brand worth paying attention to. It’s understanding that efficiency and artistry aren’t the same thing, AND that the market rewards artistry far more than it rewards efficiency.
The tools will keep evolving. The capabilities will keep expanding. But the fundamental truth isn’t going to change: people connect with people. Stories resonate because humans crafted them with intention. Brands build loyalty through authentic relationship, not optimized output.
AI is the most powerful addition to the storytelling toolkit I’ve seen in my career. But it’s an addition to the toolkit—not a replacement for the craftsperson holding it.
At Vnaqlr, we’re using every tool available to deliver exceptional work for our clients. And we’re using human judgment to know when to put the hammer down and pick up something sharper.
That’s the sweet spot. And I don’t think it’s going anywhere.
Ready to bring genuine human creativity to your marketing—enhanced by the best of what AI offers?
I am looking forward to chatting with you about your latest challenges, and solving them with the hammer AND the glass cutter.
