AI Adoption in Marketing: Building Frameworks and Avoiding Hype

AI Is Everywhere, but AI Adoption in Marketing Is Still Messy

Like crypto was 10 years ago, AI has become the subject of most conversations; from marketing and advertising to retail and logistics.

Arguably, hedging your bets with AI has done more than crypto has with all of its volatility, but it’s not without its share of controversy. Which means many topics, many opposing opinions, and many, many theories, frameworks, and yes, schemes.

As Danny Gavin put it during his presentation at the American Marketing Association (AWA), back in Q4 of 2025

“It’s late. You’re scrolling LinkedIn one last time before bed. First post: someone’s built a 50-step AI process to replace your entire marketing team. Next: a recent grad who can’t find a job in the same industry. Third: a tactic from 2015, posted like it’s news. Three posts. Three completely different takes on AI. Zero clarity.”

That’s where most marketing teams are right now.

Surrounded by noise, under pressure to do something with AI, and not entirely sure what that something actually looks like. The gap between the hype and the reality has never been wider.

Why AI Adoption Is a Cultural Challenge First

Before you can adopt AI, you have to deal with something messier: the way your team feels about it.

Resistance to AI isn’t irrational. It comes from real places. Most people don’t like change.

Uncertainty about what it means for your role, the challenging learning curve, and the overwhelm at the pace things are changing. 

It’s the quiet fear that if you don’t figure this out fast enough, you’ll be left behind.

Here’s what Danny kept coming back to in the AMA talk:

“…for most agencies, the biggest obstacle isn’t access to AI. It’s time. Agencies are built on tracked hours. Every minute is accounted for. So where does learning fit? Where does experimentation fit? If there’s no room for either, the technology doesn’t matter.”

That’s the necessary shift: “It’s not really about AI. It’s about cultivating a culture.”

AI works best when it’s positioned as something the team is building together, not something being handed down from the top.

Leadership has to create the conditions that instills trust that experimenting is okay. Otherwise, from your team’s perspective, you’re telling them to adopt technology that will replace them.

Building an AI-Positive Culture (Without Forcing It)

1. Lower the Bar – Seriously

The fastest way to kill momentum is to make people feel like their AI experiment has to be flawless before it counts. It doesn’t.

The assumption is that AI should deliver perfect results immediately. Chasing perfection from the start wastes time and momentum.

During the AMA presentation, Danny explained that their team once tried to build a fully automated system for client decks. They tried to build it perfectly multiple times.

It didn’t work.

So they did something smarter.

They built a deep research prompt in ChatGPT, took that output, and manually fed it into Gamma to create the deck. It wasn’t seamless. It wasn’t the automation we originally envisioned. But it cut the prep work dramatically, and it gave the team something real to bring into client meetings.

“A 6 out of 10 is going to be totally fine.” Danny explained.

That’s the mindset. Small improvements count. Imperfect solutions count. And going overboard is a sure-fire way to waste resources chasing the perfect automation.

2. Make Experimentation Official

Curiosity alone won’t get you there. If your team feels like AI work is something they have to sneak in on their own time, it’s not going to stick. You have to build the permission into the structure.

At Optidge, we did three things:

• We added AI work as a trackable category in our project management software, on equal footing with billable client work. 
• We created a dedicated AI task force to make it a real initiative. 
• And we told the team plainly: this is part of your job now.

Psychological safety isn’t just a mindset, it’s a system. 

Build the system, and the experimentation follows.

3. Let the Curious People Lead

Not everyone is going to be excited about AI. You don’t need everyone excited. You need a few people who are naturally drawn to it, and then you get out of their way.

Their excitement will organically influence your team. “Hey, I ditched the spreadsheet data entry and I automated it. Here, let me set it up for you.”

That’s how it actually spreads. Not top-down, but peer to peer. Find those people, give them room to experiment, and let their wins do the convincing.

The Quick Wins Playbook: Where AI Creates Immediate Value

If culture gets you started, then quick wins keep you going.

When a team sees a real, tangible improvement from a small experiment, the conversation shifts from “Should we try AI?” to “What else can we do?”

At Optidge, we used a simple process:

• Team members identified repetitive tasks and documented the step-by-step process
• Everyone got four weeks to experiment with tools and approaches
• We checked in bi-weekly to see what was working
• Whatever showed promise got refined. Whatever worked got shared across the team

Here’s what came out of it.

Speeding up Research and Analysis

AI didn’t replace research. It compressed it. That’s the difference between AI hype and AI value. 

By using AI to strategically cut down research time to focus more on human-led evaluation and improvements, analysis keeps its high value without the time commitment.

Improving Internal Documentation and Workflows

NotebookLM Automation
Team communication is scattered everywhere; meeting notes in email, updates in Slack, decisions in DMs. NotebookLM pulls it all together automatically and creates one searchable source of truth.

NotebookLM for every client: All meeting notes and brand guidelines in one place. Everything is automatically captured and organized.

Google Click ID Automation

Some workflows are so repetitive and manual that they’re basically screaming for automation. For example: manually pulling Google Click IDs and matching them to UTM parameters. 

It was tedious work that ate up serious time every single day.

Automate your Click ID process: 1.5 hours daily down to 15 minutes with a Zapier flow. Over an hour of someone’s day, every day, reclaimed!

Supporting Campaign Ideation and Iteration

Custom GPTs turned out to be one of the biggest wins:

• Ad script prompting
• Client weekly summary emails
• Paid social proposal drafts
• Timeline builders
• Resume and portfolio reviews (education side)

Each one was built around a process we already had. We just turned it into a tool anyone on the team could use consistently.

Enhancing Reporting, QA, and Process Efficiency

Danny’s email problem: between meetings, he’d open his inbox, get pulled into threads, and suddenly be late for the next call.

Solution: A Zapier flow that reads every incoming email, categorizes it by theme and priority, and sends high-priority items as a text to his phone. Built it in one evening. Still runs every day.

The point isn’t the specific tool. It’s the pattern: identify what’s eating your time, document it, and find the smallest possible solution.

From Test to Transformation: Scaling What Works

One thing we teach clients when we work to customize and improve their Hubspot platform experience is that it starts at testing. 

A quick win that stays with one person is a shortcut. A quick win that gets shared, documented, measured, and repeated is the beginning of transformation.

At Optidge, the next step after finding what worked was making it work for everyone.

That meant three things:

Standardize the Wins

Be the change and talk about it. 

The value multiplies when you pull it out of one person’s workflow and turn it into something the whole team uses. Document how it works. Share it. Make it the default.

Keep Humans in the Loop

AI gets things wrong. Danny caught an error in a report where the numbers didn’t add up. The tool had confidently produced the wrong figure. If no one had checked, that number would have gone out to a client.

Automation isn’t a reason to stop paying attention. It’s a reason to pay smarter attention. Build quality checks into every process from the start.

Build an AI Policy

This came up organically. During the AMA session, Danny mentioned when a Texas Tech graduate student interviewing him for her thesis asked: “What’s your AI policy?” He didn’t have a formal answer.

But the question stuck. 

Once AI is woven into your workflows, you need clear guidelines: how it’s used, where it’s not, how it varies by client (especially in regulated industries like healthcare).

Transformation is incremental. Test something. Validate it works. Then systematize it with the guardrails it needs to scale.

What This Means for Agencies and Marketing Teams Today

When you build AI adoption around culture and quick wins instead of big rollouts, the results show up in places you might not expect:

• Operationally, you get time back. Real, measurable hours that used to disappear into repetitive tasks now go to work that actually requires human thinking.
• Your team feels more confident. Not because they’ve mastered AI, but because they’ve built something that works. Watching your own experiment solve a problem sticks differently than being told to use a new tool.
• Your clients get better work. Faster research. More consistent reporting. Smoother transitions when teams change. These improvements compound quietly and matter a lot over time.
• Long-term, you build a competitive advantage. Not from using the most advanced AI, but from building a culture where experimentation is normal, wins get shared, and the whole organization gets better together.

Progress Over Perfection

AI isn’t going anywhere. Neither is the pressure to figure it out quickly. The good news: you don’t need to figure it all out at once.

You need momentum. One small experiment that works better than what you had. One win that gets shared. One process that saves someone’s time and makes them curious about what else is possible.

Danny closed the AMA talk with a challenge:

• This Sunday, find an hour
• Think about the task that’s been driving you crazy
• Write down how it works, step by step
• Don’t worry about making the solution perfect. A 6 out of 10 is a starting point

That’s where it begins. Not with a 50-step process. Not with a big announcement. Just one thing that works a little better than it did before.

And if you’re ready to turn your own experiments into something repeatable, that’s exactly what we do at Optidge. Schedule a time to connect with us today.