Artificial intelligence sits at the center of every marketing forecast deck, yet most teams still wrestle with basic questions: What should we automate first, what belongs on ice, and how do we keep the work human?

We know AI should make things more efficient and make our lives easier, but we’re still not quite sure how to make that happen. No one wants to look uninformed, yet few organizations have the connective tissue to turn scattered tools into concrete results—revenue, retention, or reputation.

If that scene feels familiar, this article is for you.

Real Talk: We’re Stuck in AI's Messy Middle

Artificial intelligence is both gift and threat. It can draft a press release in 30 seconds, yet it also tags the creative team with a digital expiration date. The novelty stage is somewhat behind us, full integration lies ahead, and today we navigate a construction zone of half‑finished bridges. Terms overlap, platforms over‑promise, and the path to revenue stays hazy. It’s very clear to me that we’re in the era of the messy middle of artificial intelligence.

I saw this tension firsthand at a recent Webflow executive dinner. Around the table, marketers peppered Linda Tong, CEO of Webflow, with questions like: Which tool is best? How do we keep proprietary data from leaking through prompts, and who owns the output? What guardrails stop brand damage before it spirals? And how will our teams pick up the skills fast enough to stay useful? Their unease mirrored my own. I believe in AI’s upside, yet I also recognize that my role could disappear faster than I would prefer.

Pulse Check: How Marketing Teams Really Feel About AI

Walk the halls of any marketing department and you’ll hear the same two notes playing at once: a buzz of curiosity underscored by a low, unbroken rumble of doubt. Curiosity thrives because the tools feel limitless—prompt libraries in shared drives, browser extensions that write ad copy on the fly, internal Slack channels buzzing with "what if" experiments. Yet adoption remains lumpy. One product squad may run weekly prompt‑engineering stand‑ups while the adjacent team still clings to 2015 playbooks.

Leaders no longer fear the technology itself; they fear a colleague across the aisle who figures it out first.

"Machines won’t replace us. Marketers who master machines will."

That unevenness fuels quiet trepidation. Revenue targets climb even as head‑count freezes, and every budget line now competes with the promise of "doing more with fewer humans." Leaders no longer fear the technology itself; they fear a colleague across the aisle who figures it out first.

Against this backdrop, the industry behaves like an unregulated playground. There are no settled best practices, only public experiments that sometimes blossom into case studies and sometimes implode in brand‑safety headlines. The stakes feel especially high around data privacy, IP ownership, and creative integrity. Most teams agree that AI is the new electricity—indispensable yet dangerous when wiring is sloppy.

Most teams agree that AI is the new electricity—indispensable yet dangerous when wiring is sloppy.

While curiosity sits at the top of the emotional stack, practical risks surface in five predictable buckets:

1. Mediocrity creep: brand voice flattens, eroding premium pricing power.
2. Service commoditization: differentiation shifts upstream to strategy and story.
3. Empathy gap: sentiment scores miss nuance found only in live conversations.
4. Workforce disruption: 53 percent of marketers expect a net loss of roles within three years, igniting morale dips and stealth layoffs.
5. Governance risk: unclear rules around licensing, data provenance, and regulatory compliance keep legal teams on edge.

None of these issues spell doom, but each one penalizes passivity. The marketers who learn openly, iterate in public, and document results are setting the pace for the entire profession.

Make It Real: Implementing AI in Your Marketing Ecosystem

A 2025 Edelman study shows that sixty‑eight percent of customers feel more comfortable when brands admit where algorithms lend a hand. Concealment breeds mistrust, so I will spell out exactly how we use AI today.

Intern‑Level Multipliers

We ask language models to create rough‑draft headlines, summarize calls, and propose subject lines. These outputs behave like smart interns—fast at first pass yet still reliant on human context. The payoff is time reclaimed, not talent replaced.

Automation as Connective Tissue

Single‑purpose prompts quickly turn into silos. I’m not sure about you, but I don’t want to live in ChatGPT all day. I want it to work for me. To do this, we’re playing with quite a few automations that make marketing far more efficient. Some of the most recent scenarios have been:

• Automatically create social post content from blog posts, captured in a single location that makes it easy to get into various channels.
• We are able to send natural text messages to a slack channel and have it sorted and organized in Asana, tagged correctly to the active project and correct assigned individual.
• Identify out of date blog content and provide insights on key ways to update and make relevant to our ICP.

The system is far from perfect, yet each week everything gets smoother.

Deeper Audience Insights

In lieu of first-person research, we use more advanced AI models to deepen our understanding of audience mindsets and behaviors, helping improve website user experience and make content more engaging and more effective.

Tip: Rather than focusing first on the persona, start with a general research prompt to capture industry data and pain points. Then feed those results in to help round out your audience personas.

Turning Analytics Into Digestible Insight

We're using AI to surface patterns and outliers in our analytics in minutes—not days. ChatGPT helps us maximize our tech stack, while we integrate data across platforms to create a single, richer view of performance.

AI Search

ChatGPT now sits on my phone where Chrome once lived. That personal shift reflects a profound transformation in how we search. Traditional engines are losing ground to AI systems that prioritize direct answers over blue links, context over keywords, and conversation over queries. AI search favors structured, scannable content: schema, clear headings, and concise answers. Technical SEO isn’t obsolete—it’s your shortcut into the AI layer of search.

My two cents

AI is neither savior nor saboteur; it is leverage. Use it early, measure honestly, and iterate in public.

And as Linda suggested at that Webflow dinner, identify the tasks that slow you down, hand them to the machine, and reinvest the reclaimed hours in decisions only humans can make. Stack a dozen of those micro-reliefs and the cumulative effect feels like a breakthrough—minus the hype and the budget shock.

Small, repeatable wins add up faster than grand unveilings. Every prompt we refine, every micro-automation we wire, and every schema tweak we publish moves us one notch closer to an organization where people focus on thinking and machines handle the tedium. That’s the spirit behind the simple rule set I lean on—start before certainty, keep humans in the loop, measure what matters, and share what works. Hold those four and you’re ready for the next, final step.