CXL - AI in marketing - "It's not there yet"

"It's not there yet" echoes around the table when senior marketers discuss AI.
Opinion piece by Stefan Maritz, from CXL

Open LinkedIn. Scroll for 30 seconds. You’ll see a flood of AI screenshots, “10x productivity” threads, and founders bragging about agents replacing teams.

If you believe everything you see on LinkedIn, it looks like everyone’s crushing it.

But the truth is, most people are just winging it. The shiny object syndrome has never been this bad. Everyone’s jumping on everything, but very few know what the hell they’re doing. They saw a demo, copied a prompt, and now they think they’re experts.

When you talk to senior marketers investing the time themselves - trying, testing, building systems - you’ll hear the same line over and over: “It’s not there yet.”

They want the tools to deliver more. They’re getting glimpses of value, but keep hitting walls. Everyone’s chasing magic. Few are getting results they really want.

I think it's important that we stay grounded throughout this, we are slap-bang in the middle of an industrial revolution after all. Let's talk about the current state of things without getting carried away.

1. Shiny tools, shaky execution

According to HubSpot’s 2024 report, AI adoption among marketers jumped from 21% to 74% in just one year. That’s a massive leap. But don’t confuse adoption with effectiveness.

What most people call “using AI” is just surface-level prompt dumping. They generate a caption or rewrite a blog intro and call it a day. That’s not operational transformation. That’s novelty use.

It’s like handing a scalpel to someone who’s never done surgery and asking them to perform. Just because a tool is powerful doesn’t mean it’ll be used right. Most aren’t training it, iterating, or applying it with a clear system. They’re just reacting to the noise.

The result? A sea of generic outputs, tone mismatches, and bad brand copy-pasted from GPT.

2. The gap between expectation and reality

The biggest letdown isn’t the tools - it’s the expectations people put on them.

People want AI to replace their thinking. They expect it to come up with strategy, tone, and differentiation. That’s not how it works. At this very specific point in time, AI is a multiplier - it amplifies what you feed it. Garbage in, garbage out. Vague prompts get vague outputs. Lazy inputs get lazy "thinking".

Strategy, critical thinking and intuition still need to come from you as a human being with an actual brain. Positioning still needs human insight. Messaging still needs sharp judgment.

You can’t outsource your core thinking and expect to get gold back. AI can assist, not lead.

Think of it like a co-pilot. You still have to fly the damn plane.

3. AI is a companion, not a shortcut

If you want to get actual value from AI, treat it like a high-potential junior team member. It’s not the magic box that spits out genius. It’s a collaborator.

Open it in the morning. Keep it running all day. Make it part of your workflow. Get comfortable speaking to it, delegating to it and reviewing its output like you would do with a junior. Let it assist in research, data analysis, idea generation, rewrite your briefs, pressure test your ideas, and tighten your arguments.

Give it your tone of voice. Feed it customer pain points. Show it your best performing ads and landing pages. Keep training it over time.

When you ask it for work, critique what it gives back. Push it. Iterate. Guide. That’s when it gets better. That’s when it becomes leverage.

If you’re just using it for quick content and shipping as-is, you’re missing the point, and quite frankly, probably shooting yourself in the foot. You will go backwards faster than anything else.

4. No process = no progress

What separates tourists from operators? Operators build systems.

The smartest marketers experimenting with AI right now are documenting everything. Best prompts, working outputs, failures, learnings. They’re building internal knowledge bases. They’re versioning GPTs. They’re creating SOPs that let their workflows scale.

This isn’t about tools - it’s about compounding knowledge.

“AI experience” is already showing up in job descriptions EVERYWHERE. In a year, it’ll be like knowing how to use Google Workspace or Slack. Table stakes. You either have it or you don’t.

If you’re not documenting your AI playbook as you go, you’re going to get left behind.

5. The marketer who knows AI will replace the one who doesn’t

The question isn’t “Will AI replace me?” It’s “Will I be outcompeted by someone who uses it better?”

The answer: probably, yes.

Companies aren’t looking for marketers who know how to click around ChatGPT or ask it how to write a creative brief. They’re looking for people who bring actual leverage - custom GPTs, refined workflows, time-saving automations, quality-controlled output.

They want someone who can bring their own stack to the job. Someone who’s figured out how to use tools to speed up customer research, synthesize market insights, and generate high-quality briefs in minutes.

You’re not just bringing skills. You’re bringing a system.

6. What to do right now

If you’re not already working like this, it’s time to change how you operate. Here’s how:

1. Play every day
Test something new every week. Try custom GPTs. Experiment with audio and video tools. Push the edges. 99% of people still haven’t gone beyond basic prompt use. Be the 1%.

2. Create custom workflows
Build a system that mirrors how you work. Set up reusable prompt stacks. Embed your brand tone. Save your templates. Create repeatable processes. Make it yours.

3. Work in public
Don’t hoard your experiments. Share what’s working. Share your findings with your team and network.

4. Document like it’s SOP
Every time you land on something that saves time or improves quality, turn it into a system. Capture the prompt. Add notes. Save the output. Share it with your team.

5. Treat output as draft, not final
AI gives you raw material. You are the editor. You are the quality control. Never publish straight from the model. Shape it. Tighten it. Make it yours.

This is the time to build

AI isn’t mature yet. It’s inconsistent. It breaks. It gives B+ work at best.

But that’s the opportunity.

This is when reputations are built. While most marketers are overwhelmed or lost in tool overload, you can carve out an edge by doing the unsexy work - testing, training, systemizing.

Build your stack. Refine your approach. Document everything. Be an early adopter. The next 12 months will separate the marketers who just added “GPT” to their headline from the ones who actually built capability.

Everyone’s watching the parade. Be the one driving the float.

CXL-trained marketers get hired faster, promoted sooner, and paid more because they know how to test, optimize, and scale what works. 

• 120+ Practical marketing courses
• 9 Minidegree programs to build deep skills
• Learn directly from leading practitioners
• Apply what you learn immediately

Learn more >

CXL, 901 S Mo Pac Expy, Suite 150, Barton Oaks Plaza One, Austin TX 78746.

Follow us on LinkedIn.

Unsubscribe from future emails like these.