• Everyday AI
  • Posts
  • Ep 722: How to Build a Team of AI-savvy Marketers

Ep 722: How to Build a Team of AI-savvy Marketers

Nano Banana 2 drops, AI leaders to meet with Trump, Perplexity launches safer ‘Computer’ Agent, Claude Cowork launches scheduled tasks and more

Author

Everyday AI
February 26, 2026

Sup y’all! 👋

Did you miss this from yesterday?

Welp, you might wanna read this then.

As part of research for yesterday’s show on amazing new NotebookLM updates, we created a simple, done-for-you workflow where you can easily convert NotebookLM’s amazing horizontal slides into share-worthy mobile-first PDF carousels.

All you gotta do is repost yesterday’s NotebookLM show on LinkedIn and we’ll send you the simple instructions, video walkthrough and our custom-built GPT below that does the heavy lifting.

(You know if I’m sharing this twice, it’s REALLY, REALLY good)

✌️

Jordan

Outsmart The Future

Today in Everyday AI
8 minute read

🎙 Daily Podcast Episode: How can you build an AI-savvy marketing team when AI is reshaping marketing daily? Give today’s show a watch/read/listen to find out.

🕵️‍♂️ Fresh Finds: Toy Story takes aim at creepy AI, Cursor’s big agent update, Nano Banana 2 benchmarks and more.  Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: Nano Banana 2 drops, AI leaders to meet with Trump, Perplexity launches safer ‘Computer’ Agent, Claude Cowork launches scheduled tasks and more. Read on for Byte Sized News.

💪 Leverage AI: We tapped into the brains of a marketing leader to figure out how to build AI-savvy marketing teams, even when AI is reshaping the landscape weekly. We drop the blueprint. Keep reading for that!

↩️ Don’t miss out: Miss our last newsletter? We covered: Report says The Pentagon ditches Claude for Elon's Grok, Anthropic brings remote control to Claude Code, OpenAI drops GPT-5.3 Codex in the API and more.   Check it here!

Ep 720: China Stealing AI from the U.S.? Inside Anthropic’s Bombshell Allegations


Facts: AI started mainly as a marketing tool. 📢

Sure, many marketers have been ahead of the curve when it comes to AI tools and creating high-quality, AI-driven outputs. 

But now they face a major problem: decades of marketing experience can no (largely) be achieved with a few reps of today's latest AI tools. 

So, what's your company supposed to do? And how can you possible build a team full of AI-savvy marketers? 

Also on the pod today:

• Building an AI council 🧑‍💼
• Ditching old AI platforms fast 🏃‍♂️
• “Middle to middle” AI workflows ↔️

It’ll be worth your 27 minutes:

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – KoiDex heps you know if an AI model is safe, Rover lets you turn your website into an AI agent, Tessl helps you evaluate agent skills.

 

OpenAI vs. Anthropic — Here’s more context on the ongoing battle between the two AI heavyweights.

Nano Banana 2 Benchmarks — Nano Banana 2 is the best AI image model available now, according to Arena scores.

AI in Movies -- Toy Story 5 is taking aim at creepy AI toys.

xAI Complaints — Elon Musk’s xAI turbines are shaking up Southaven, Mississippi, with nonstop noise and pollution sparking fierce community pushback. What’s the real cost of powering AI?

AI and Politics — Here’s why Florida Gov. DeSantis is going anti-AI, even though it goes against others leaders in the Republican party.

Cursor Updates — Cursor’s new cloud agents now run in isolated VMs, fully build and test code, and submit merge-ready pull requests with video proof.

AI Acquisition — Anthropic scoops up Vercept, shuts its product, and investor drama unfolds. What’s next in the AI showdown?

1. AI Leaders to Meet Trump on Power Pledge for Data Centers 🇺🇸

Next week, CEOs from leading AI companies including Amazon, Google, Meta, Microsoft, xAI, Oracle, and OpenAI will convene at the White House with President Donald Trump to sign a pledge ensuring their AI data centers supply their own electricity.

This timely move aims to address public concerns about rising energy costs linked to AI infrastructure growth. Trump emphasized the companies’ responsibility to build or secure their own power sources to avoid burdening consumers with higher bills.

2. Anthropic Launches Claude Cowork with Task Scheduling for Workers 👷

Anthropic has updated Claude Cowork to let users schedule and automate tasks like managing files and researching info. The AI runs safely in a virtual machine and needs the Claude Desktop app open on an awake computer to work.

Users can set recurring tasks through a simple schedule feature and control what the AI can access. This makes powerful AI help easier and more practical for everyday work.

3. DeepSeek Keeps U.S. Chipmakers in the Dark on New AI Model 🐋

Chinese AI lab DeepSeek is breaking industry norms by withholding its upcoming V4 model from leading U.S. chipmakers Nvidia and AMD, instead giving early access to domestic firms like Huawei. This move comes amid rising tensions and appears to be part of China’s strategy to limit U.S. hardware advantages in AI development.

Despite prior collaboration with Nvidia, DeepSeek is prioritizing local suppliers ahead of its anticipated launch. The development also raises questions over possible use of U.S. chips in China under export control scrutiny, highlighting growing geopolitical complexities in AI innovation.

4. OpenAI Recruits Meta AI Chief Who Reportedly Had a $200M+ Package at Meta 🤑

According to The Information, OpenAI has hired Ruoming Pang, who oversaw AI infrastructure at Meta’s Superintelligence Labs and left the company last week after joining Meta from Apple about seven months ago.

Pang had joined Meta on a compensation package valued at more than $200 million over several years, underscoring the high stakes in the current Silicon Valley talent war. Meanwhile, Apple continues to lose AI researchers to Meta and Google DeepMind, reflecting internal challenges and dissatisfaction over outsourcing AI development. These moves highlight the intense competition among tech giants to secure top talent and lead the next wave of artificial intelligence innovation.

5. Nvidia posts impressive earnings but stock slips on AI spending worries 📈

Nvidia reported fiscal Q4 revenue of $68.13 billion, beating expectations and driven by data center sales that made up 91 percent of revenue, yet the stock fell about 3 percent as investors questioned whether AI infrastructure spending can stay that strong.

Analysts and portfolio managers say the market is now more worried about the sustainability and monetization of AI capex and the cashflow pressure on hyperscalers that have been the company’s biggest customers. Management’s bullish guidance for roughly $78 billion in next-quarter revenue surprised the street and suggests growth may continue, but it also sharpened the debate over how long hyperscalers can keep funding this pace

6. Perplexity launches "Computer" multiagent AI worker 🖥️

Perplexity launched Computer, a multiagent system for safer autonomous desktop assistants, now available to Perplexity Max users and soon to Pro and Enterprise. It coordinates over a dozen specialist models—Claude Opus 4.6 for reasoning, GPT-5.2 for recall and search, plus others for imagery and light tasks—to handle complex workflows.

Computer aims to be safer than always-on agents by running tasks in a secure sandbox and letting users control model assignments. This reflects growing interest in autonomous agents while addressing security risks after recent unpredictable agent incidents.

6. Google unveils Nano Banana 2 for higher-fidelity, faster image generation 🍌

Today Google DeepMind introduced Nano Banana 2 (Gemini 3.1 Flash Image), a faster Flash image model that boosts visual fidelity, text rendering and localization, and adds new aspect ratios and a 512px tier to speed iterations.

The update leverages Gemini’s world knowledge and web image search for more grounded visuals, and offers configurable "thinking" levels so developers can trade speed for deeper reasoning.

Your marketing team prolly cranked out 10X more content with AI this year compared to a few years ago.

So did everyone else's.

(And sorry…. but when everyone's faster, speed ain't the advantage anymore.)

The real gap forming in marketing right now isn't between teams that use AI and teams that don't. It's between teams that tied AI to actual revenue and teams that are still high-fiving over blog post volume.

One CMO managing 130 marketers just walked us through the exact structural playbook that actually makes a difference in helping build AI-savvy marketing teams.

We unpacked it all on today's Everyday AI. AI councils, a "middle to middle" framework most teams have never heard of, and net new roles that didn't exist 24 months ago.

Let’s get savvy shorties.

1. Stop Counting Content Start Counting Revenue 🔥

Scott Morris, CMO of Sprout Social, walked into 130 marketers all experimenting with AI in their own little bubbles.

The output went up. Nobody could tie any of it to pipeline.

So he stood up an AI council. A small crew of team leads with one job. Define clear business outcomes, track what's actually working, and kill what's noise.

The shift that mattered wasn't better tools. It was moving from "everyone do your thing" to structured workflows with brand voice baked into the AI systems themselves.

Most teams skip this entirely. They buy seats, hand them out, and wonder why the output all sounds like it was written by the same robot.

Scott's team went from measuring volume to measuring conversion rates, trial-to-paid velocity, and cross-functional impact with sales and product. That's the difference between subsidizing expensive autocomplete and building something that compounds.

Try This

Grab two or three team leads this week and give them one mandate. Define three to five business outcomes you actually want AI to move. Pipeline. Conversion. Customer experience. Not "produce more."

Meet biweekly. Surface wins and kill what's noise. Our model has always been measure what matters and slowly sprint.

2. AI Is Middle to Middle Not End to End ⚡

Scott said his team’s biggest early mistake was assuming AI was a complete solution.

You know… hand marketers a tool, get polished work, ship it.

He learned it's middle to middle. The front end and the back end are still 100% human jobs.

(We call this the Human AI Hamburger)

If your team isn't trained on prompting, they're feeding AI garbage and expecting gourmet meals. And if nobody's checking output against brand voice, everything sounds like it was assembled by a committee that all read the same LinkedIn posts.

The fix was encoding brand voice into their AI tools and building structure around both ends of every interaction. The marketers who master both sides produce work that sounds human. The rest are straight up contributing to the content landfill and don't even know it.

Try This

Take your brand voice guidelines and turn them into a custom system prompt this week. Gemini users, build a gem. ChatGPT folks, create a custom GPT.

Run a blind test. Have three people process the same brief with and without the brand encoding.

The gap will be ugly. That gap is exactly what your audience sees every time your team skips this step. Make the encoded version the default before another piece goes live.

3. Create Roles Nobody Is Hiring For 🚀

While most companies ask how to train current marketers on AI, Scott asked a different question. What roles do we need that nobody has ever hired for?

He built a Senior AI and Creative Technology Strategist. One person whose entire job is evaluating tools, testing integrations, and building repeatable systems the whole team taps into.

Here's why this matters.

Some tools Scott onboarded two years ago are already obsolete because platforms like Gemini caught up. If nobody on your team has "think six months ahead on AI" as their actual job description, you're just gonna keep reacting to whatever drops next Tuesday.

The companies that create these roles first are gonna be the ones everyone else copies in 18 months. That pattern has played out with every major technology shift. This one's moving faster.

Try This

Designate one person as your AI workflow lead this week. Give them four to six hours for tool evaluation and integration testing.

Have them document and present findings monthly so the whole team benefits.

Start writing the job description for a dedicated AI marketing strategist now. This is one of those moves that looks optional today and wildly obvious by Q3.

 

 🚨 Bonus 🚨

Did you miss this from yesterday?

Welp, you might wanna read this then.

As part of research for yesterday’s show on amazing new NotebookLM updates, we created a simple, done-for-you workflow where you can easily convert NotebookLM’s amazing horizontal slides into share-worthy mobile-first PDF carousels.

All you gotta do is repost yesterday’s NotebookLM show on LinkedIn and we’ll send you the simple instructions, video walkthrough and our custom-built GPT below that does the heavy lifting.

Reply

or to participate.