The Marketing AI Conference in Cleveland featured an impressive lineup of speakers who posed big questions we’re still thinking about, like how roles will evolve as we race ahead toward artificial general intelligence (if we’re not already there) and what work is too important to outsource.
Led by SmarterX and Marketing AI Institute founder Paul Roetzer, it was a solid mix of inspirational keynotes and practical sessions. While there were plenty of tutorials on how to set up better skills, prompts and workflows, what stood out to me was how much of the conversation focused on the uniquely human work that deserves more of our attention.
Roetzer argued there are three main roles that will dominate the world of work: The architect, who designs how things get done; the orchestrator, who directs people and agents and evaluates the output; and the apprentice who builds judgment and taste.
That judgment layer is more important than ever when AI is handling so much work that was traditionally done by entry-level people who learned from it. It was refreshing to hear others say that while AI has saved them many hours on some tasks, they’re busier and more overwhelmed than ever. When anything can be done in a lot less time, we start to ask ourselves what’s still worth doing.
And how do we continue advancing at this pace without losing sight of what matters?
Here are six themes we took away from the conference that will inform our priorities (and yours) in the coming year/s.
There are times when I wonder, “Is anyone even reading this?”
It’s a fair question. Author Joe Pulizzi, who helped define content marketing during the early days of educational blogs and gated whitepapers, argued we’ve surpassed a critical mass of internet saturation. About three-fourths of all website content is at least partially AI-generated, and AI is responsible for about half of all new books published on Amazon, according to the company’s own data.
Trusted human voices are becoming more rare and valuable.
People are talking about your brand whether you invite them to or not, he said, as Chick-fil-A discovered when one of their employees started posting viral TikTok videos of her menu hacks. When they asked her to stop, she got a deal as a paid influencer for Shake Shack instead.
To stand out today, he said, your marketing team needs to think of themselves as franchise owners instead of just content producers.
As Pulizzi put it, Saturday Night Live is the brand, but people tune in to watch Weekend Update (and my newer favorite, Marcello Hernandez).
That means your job is to be Lorne Michaels. Set the guardrails for what they create and make sure you’re distributing it and rewarding it when it’s successful. There’s a place for everyone, whether they’re an expert able to explain complex ideas, a storyteller or a leader who inspires.
All this contributes to a strong brand experience, something you can’t outsource to AI. It’s the feeling someone gets when they interact with your company and what nudges them to choose you over anyone else.
You can let AI surface patterns, draft opinions and pressure-test ideas, but the team still needs to make the calls that define your brand: everything from deciding what headline to use to figuring out why a website page isn’t converting to what turns customers into lifelong brand ambassadors.
Your next move: Figure out where trust is lacking among your customers or where something could be explained in a clearer, more relatable way. Think about who the best person to tell that story might be and what that person needs from you. Maybe it’s encouragement to capture their own videos while they’re out in the field or a monthly “ask me anything” session with your customers that could be turned into short video clips and longer explainer pieces.
Over the past year, we’ve been building a library of Claude skills for ourselves and our clients. We share projects and notes and the occasional skill if it seems useful, but we’re still working on building the engine that keeps it running and improving.
Brian Piper’s session on building a content marketing agentic workflow was the swift kick I needed to level up our team’s skills and connect them together.
While he’s developed impressive content engines that run on over a dozen different skills, he got my attention immediately by explaining how he keeps the “human in the lead” (not just in the loop).
These skills are trained on everything from industry news and a company’s personas and stories to answer engine optimization and content distribution.
The system is led by an orchestrator skill that learns from inputs and feedback, so it constantly improves the more it’s used.
For Piper, this process has compressed what used to be a four-week process into a matter of hours.
While AI can do much with the context it’s given, it will always default to whatever broader information it’s trained on, which is, in essence, the entire internet. It’s good at giving you a predictable response or summarizing something, but it’s not so great at knowing how to tell a strong story and what parts to leave out.
“Don’t let AI write the story,” Piper said. “You bring the story.”
Instead of running to AI for your first draft, give it a brief of what you want to talk about and ask it to interview you to develop a story or process. If you’re following the same process more than three times, make it a skill.
It’s just as important to decide where the human checkpoints should be and where you don’t want to rely on AI.
Amanda Todorovich, who leads content strategy at Cleveland Clinic, said her team has built more than 60 custom GPTs to help with everything from understanding the patient’s perspective to writing code. But every article is still fully written by her team and reviewed by a physician, because the stakes are justifiably high.
“We’re focusing on using AI to make content that’s better, not faster,” she said. “For instance, are we writing more comprehensive articles because our writers are using AI to ask better interview questions?”
Your next move: Audit the skills or custom GPTs you’ve built and compare them to your writing or marketing process. Look for what’s missing. Do you need a more detailed collection of customer stories? A better way to come up with ideas?
The marketer role is evolving, and developing original content that leads to stories others will share is becoming a bigger part of it.
SmarterX Director of Research Taylor Radey and Chief Content Officer Mike Kaput both shared examples of how their team uses Claude for research projects like their State of AI for Business report.
AI’s deep research feature set gives users a high-level overview of a topic and help develop questions that produce data no one else has.
In addition to fine-tuning questions, analyzing the data and distilling it into main findings, it can also help users verify the work – but it’s still important for people to do the validation work, as well. Make sure every source cited goes to a live, verifiable link and all the numbers add up.
This kind of research is irresistible to journalists hungry for statistics no one else has, particularly if they support a compelling premise or narrative.
Branded mentions are a core factor in AI visibility, so investing in public relations can pay big dividends.
Nicole Franco, Fractl’s Head of Digital PR, shared how she gets journalists to pay attention to her clients by combining AI with good instincts.
Her agency analyzed 11,000 prompts to identify the best opportunities for brand visibility in AI-generated responses and media pitches. With that data at their disposal, they developed a prompt to find news from the past few days related to topics they wanted to discuss.
For one client, they uncovered a wave of searches about drinking prune juice and melted butter to lose weight. They reached out to journalists offering to connect them with a registered dietician who worked with the company. This earned them more than 20 stories that mentioned the brand.
For this to work well, of course, a newsworthy angle is needed. It should be surprising, timely, emotionally compelling, educational or entertaining, Franco said, and backed by data whenever possible.
Your next move: Think about the publications you’d love to be featured in. Analyze the stories they’ve published recently and think about what you could bring that adds to the conversation. Look for what journalists are already asking for related to this topic on Help A Reporter Out (HARO) or Qwoted. Then reach out, offering your expert to be interviewed about it.
Dale Bertrand, SEO Strategist and founder of Fire&Spark, made the case for a different kind of visibility audit.
Succeeding in search today is no longer just about ranking for specific keywords, but understanding whether your content answers the questions people are asking AI.
He recommended building a taxonomy for your industry and what topics you want to offer expertise on.
Analyze other highly visible pages in an industry similar to yours to determine what they include. You can use Schema.org to determine the entities that map to each attribute, such as a specific service or broader topic. Then, build a Claude skill that analyzes your own pages to determine what you cover and what you’re missing.
For instance, many pages leave out details about pricing, specific services or credentials. Once you add that information to your page, you can add schema markup/structured data (the preferred language for both Google and answer engines) so it’s easily searchable.
You can even build a skill to send you real-time alerts for certain metrics, such as when a page has a high number of views but low engagement or conversions.
This was encouraging to see because it’s similar to how our SEO team works to find the gaps on our clients’ websites and address them.
HubSpot’s AEO grader is a good starting point for getting suggested prompts based on what you offer and what searches are showing up in large language models. But unlike Google’s search engine results page where results are somewhat stable and predictable, AI search will give you different results each time, and each engine is different. Referrals from ChatGPT are still dominating, but other models like Claude and Perplexity are catching up. That’s why it’s important to constantly test and monitor. You can’t run an audit like this once and assume the results will be the same months later.
Your next move: Run an AEO entity gap audit. Pull your top service pages and compare what they cover to the questions people are actively asking AI tools in your category. The gaps are usually obvious and something you can fix.
It’s easier to be AI native when you’re a solo entrepreneur or a startup. How does a midsized or larger team drive AI adoption and streamline their processes?
As CEO of Good Karma Sports CEO Craig Karmazin shared, it doesn’t happen without some friction.
The team rolled out AI licenses across his 500-person company in waves of 40 at a time. He built an AI council, runs biweekly training calls back to basics, and created three career paths that didn't exist before. When someone leaves the company now, he treats the opening as a chance to grow existing teammates.
“Our teammates aren’t scared of this,” he said. “They’re realizing this is the biggest opportunity they have in their careers.”
One of the main ways they’ve used AI to innovate was in building their own digital billing and planning system that works better than anything they’d tried before.
They’re moving away from Excel spreadsheets using projects within large language models and templatizing campaigns and slide decks they can build in minutes instead of hours. This allows Karmazin greater focus on the human side of managing his company, including having more face-to-face meetings with employees and partners.
Now they’re focusing on benchmarking the time and the quality of every output and redefining career descriptions as their own innovation specialists develop new ways to do things. They’ve also realized the importance of keeping and developing their own “superstars” who continue to do what they do best, with or without AI.
He ended with three solid words of advice for teams undertaking similar challenges: “Have patience, urgently."
Holly Chen, VP of Growth Marketing at Samsara, reminded us that even if we all feel like we’re already behind, no one is an “AI native.”
“Everyone is using AI, but very few teams have it integrated into their actual workflow,” she said. “The job isn't to hire someone AI-native. It's to make every marketing team AI-native."
To do this, Chen built what she calls an AI Builder Bootcamp. Each workstream nominates two people over three weeks, and everyone is required to build an agent for their own workflow. The best ones go into production. She also added a standing "AI in Action" slot to the company’s monthly all-hands meeting where each team takes a 15-minute turn to show what they built.
“Don't just make it a talent show,” she said. Ask your team members what they did that didn’t work and what they learned from it.
In three years, Chen's team went from just three percent of the marketing team building workflows and agents to ninety.
Anyone can learn the tools; the real work is in building up the champions, shared repositories, performance review integrations, and permission to experiment with turning AI adoption into lasting change.
The biggest takeaway from MAICON 2026? Adopting AI isn't the finish line. There are still great opportunities in connecting the right tools and workflows with strategies that keep people first and foremost.
Every organization is in a different place. Some need to get their AI marketing foundation right. Others are ready to put AI to work across campaigns and workflows. And some need an experienced partner to connect the puzzle pieces and foster measurable results.
At Kuno, we start where you are and build toward that connected approach together, propelling your team toward your brand and revenue goals.
If you’re not sure where to even start, there’s nothing wrong with that! Our free AI Marketing Foundation Assessment identifies where your footings are solid and where they may need work to best accommodate AI tools or integrations.