When AI chaos meets primetime 🎬

Hello marketers. Welcome to AI Marketing School, where we dish out the latest and greatest in AI-powered marketing.

In this week’s issue:

• AI Marketing Update: Three massive AI marketing stories

• The Stack: Some cool, useful, and trending AI tools

• AI Events: Our recommended AI marketing events for networking and connections

Onwards!

AI MARKETING UPDATE

This issue is a bit late — it’s been a hectic summer — but it’s been a great opportunity to catch up on what’s been going on.

There have been some truly compelling stories from the AI marketing world lately, and now that we’re beyond halfway through 2025 (eek!), we can see that AI’s role in marketing is forging ahead, both in the form of creativity and production, research, analysis, advertising, and so on.

Three massive story breakdowns for you:

Story 1: Kalshi’s $2,000 NBA commercial

Remember when Coca-Cola's AI Christmas ad got absolutely roasted last year? Or when Google's "Dear Sydney" Olympics commercial was so tone-deaf they had to pull it? 

We now have another AI advert case study, and it represents genuine progression as text-to-video grows more sophisticated. 

During Game 3 of the NBA Finals, exchange and prediction market Kalshi aired a 30-second commercial that looked pretty insane.

A farmer floating in a pool of eggs. An alien chugging beer. An older man, draped in an American flag, screaming, "Indiana gonna win baby."

Unlike Coca-Cola's generic AI Christmas imagery or Google's heartless Olympics spot, this was chaos with purpose. The main takeaway for most? It was made with AI in 2 days for $2,000…

The filmmaker behind it

PJ Accetturo, the "AI filmmaker" behind the commercial, was hired by prediction market Kalshi to create "the most unhinged NBA Finals commercial possible."

What he delivered was a chaotic, GTA-style masterpiece that proved AI can transform a concept into primetime television in 48 hours.

Accetturo understood that AI excels at unexpected, high-energy chaos – so he made that the brand strategy, not a limitation to work around. This is quite similar to Atera last year, which also embraced AI’s unpredictable aesthetic in their successful ad. 

How Kalshi created the ad

Refreshingly, Accetturo shared his entire process, and it's surprisingly straightforward:

Step 1: Scripting

He started by asking Kalshi's team for key dialogue snippets they wanted to include. Then he brainstormed "10 wild characters in unhinged situations" to deliver those lines. 

Using Gemini and ChatGPT as creative partners, he generated concepts like "rizzed out grandpa headed to the club" and "old lady in front of pickup truck that says 'fresh manatee.'"

Instead of trying to make AI sound human and emotional, he leaned into AI's strength at generating bizarre, memorable scenarios.

Step 2: Prompt engineering 

Accetturo asked Gemini to convert every shot description into detailed Google Veo 3 prompts.

His critical rule: "I always tell it to return 5 prompts at a time – any more than that and the quality starts to slip."

Each prompt had to fully describe the scene as if Veo 3 had no context of previous shots. This meant re-describing the setting, character, and tone every single time.

Here's one of his prompts: "A handheld medium-wide shot, filmed like raw street footage on a crowded Miami strip at night. An old white man in his late 60s struts confidently down the sidewalk, surrounded by tourists and clubgoers. He's grinning from ear to ear, his belly proudly sticking out from a cropped pink T-shirt..."

Step 3: The generation marathon 

He ran these prompts through Veo 3 in batches of three, using multiple browser windows simultaneously. Generate, review, adjust, repeat. It took 300-400 generations to get 15 usable clips.

This is the key difference from traditional production: when iteration is essentially free, you can afford to be extremely picky about quality.

Step 4: Post production

Final assembly in standard editing software – nothing revolutionary, just solid editing to tie the chaos into a coherent narrative. CTAs for the platform appear at the end “The world’s gone made. Trade it.”

Accetturo believes "high-dopamine Veo 3 videos will be the ad trend of 2025." Looking at the economics response, he's probably right.

There’s an interesting philosphical difference between this any say, Coca-Cola, tried to use AI to recreate traditional advertising and fell into the uncanny valley. 

For small businesses that could never afford a $400K commercial, this makes TV advertising accessible. For larger brands, it means testing dozens of creative concepts for the price of one traditional ad.

More importantly, this opens up entirely new creative territories. When iteration is cheap and fast, you can take creative risks that would be impossible with traditional production budgets.

Story 2: Reddit unlocks its content for commercial purposes

I’m surprised that more people aren’t talking about this. Reddit just launched two tools that let brands systematically use Reddit content for marketing:

• Reddit Insights analyzes Reddit's 22+ billion posts and comments to give brands insights about what people are saying. Instead of manually reading subreddits, marketers can ask questions like "What do people think about my product?" and get AI-powered answers based on actual Reddit discussions.

• Conversation Summary Add-ons put real Reddit comments directly in ads. Here's how it works: Reddit's AI identifies positive posts about a brand or product, creates a concise summary (approximately 120 characters), and displays both the summary and the actual user posts below the advertiser's creative. More from Reddit here.

Announcement 1: Reddit Insights’ social listening

Reddit Insights uses AI to analyze conversations across the entire platform and deliver easy-to-understand insights in an interface similar to ChatGPT. Some key uses for marketing:

• Competitive intelligence: View what people really think about your competitors when they think brands aren't listening

• Product development: Find unmet needs and improvement opportunities through community feedback

• Cultural forecasting: Identify trends emerging in niche communities before they hit mainstream

Announcement 2: Conversation Summary add-ons

These ad features dynamically pull positive Reddit comments and display them directly below your advertiser creative, with AI-generated summaries of community sentiment.

For example, if Toyota runs an ad for their new SUV, actual Reddit comments might appear below like "This SUV is amazing, had mine for 2 years" alongside an AI summary saying "Redditors praise this SUV's reliability and performance."

As one commenter said “Reddit’s now packaging your hot takes and shower thoughts as marketing gold, slapping them under ads like “real talk, real results.” It’s social proof at scale: powered by AI, monetized by vibes. Great for brands, weird for users. Welcome to the age of accidental influencers.”

Early testing showed Conversation Summary Add-ons achieved 19% higher click-through rates compared to standard image ads. 

Publicis Media Exchange served as the exclusive alpha tester, collaborating with major clients such as The Hershey Company and Comcast.

For their clients, Publicis teams were able to:

• Discover new audiences they hadn't considered

• Identify specific opportunities to increase brand awareness

• Understand brand sentiment with unprecedented granularity

• Engage more strategically with Reddit communities

As Lucid's marketing VP put it: "In today's advertising world, consumers want more than polished campaigns. They're looking for validation from people like them."

Clever Reddit…

This ad format is pretty smart.

As AI-generated content floods every other platform, authentic human insights and product commentary become exponentially more valuable.

Reddit COO Jen Wong said, "In a world increasingly flattened by AI, the value of human content on our platform only grows."

Reddit reported $358.6 million in Q1 ad revenue, a 61% year-over-year increase. So they're evidently figuring out how to monetize their authentic conversations without destroying what makes them valuable.

Unlike the sanitized, brand-safe conversations on other platforms, Reddit discussions tend to be unfiltered and honest. It could backfire on some brands, even despite Reddit saying only positive mentions will be added.

At least that’s the idea. We’ll see how people game it or blame it soon enough. 

Story 3: Meta's 2026 automation plan 

Here’s something we overlooked in the ruck of news. Meta said they’re aiming for complete advertising automation by 2026.

The vision is elegantly simple and terrifying for agencies. You give Meta a product image and budget. Their AI creates the entire campaign – visuals, copy, targeting, placement, and runs it for you.

At a recent conference, Zuckerberg laid out the endgame: "We're going to get to a point where you're a business, you come to us, you tell us what your objective is, you connect to your bank account, you don't need any creative, you don't need any targeting demographic, you don't need any measurement."

Remarkably, when this news broke, marketing and advertising agency stocks dropped immediately:

• WPP: Down 2.2%

• Publicis Groupe: Down 3.8%

• Omnicom Group: Down 3.2%

• Interpublic Group: Down 1.9%

Meta is particularly focused on small and medium-sized businesses that lack dedicated marketing teams. For a local restaurant or e-commerce store, the promise of professional-grade advertising without agency fees is incredibly compelling.

However, while AI can optimize campaigns and generate creative assets, it can't develop brand strategy.

As we’ve covered many times, the best brands will hold AI’s hand rather than be led by it. 

THE STACK

Mid-year AI tool round-up

Ok time for a good old round-up of some AI marketing tools, or AI tools suitable for marketing uses. Starting with…

Midjourney V1 video

Unlike some overly complex video generators that try to do everything, Midjourney focused on doing one thing exceptionally well: turning beautiful images into short, high-quality video clips.

The workflow is elegantly simple. Create images in Midjourney as you normally would, then press "Animate" to bring them to life. You get four different 5-second video options per job, and you can extend each one by 4 seconds at a time up to 20 seconds total.

There are two animation modes:

• Automatic: AI creates motion prompts and animates automatically – surprisingly effective for ambient scenes

• Manual: You describe exactly how you want things to move and develop

Plus two motion settings:

• Low motion: Better for ambient scenes where the camera stays mostly still and subjects move deliberately

• High motion: Everything moves – both subject and camera – but higher chance of visual artifacts

Currently, video generation costs approximately eight times more than image generation, which is relatively competitive compared to other models. 

Unlike Sora, which focuses on complex physics simulation and long-form narratives, or Runway, which emphasizes cinematic realism, Midjourney doubles down on aesthetic quality and consistency.

Your videos might not show complex action sequences, but they look absolutely stunning. 

Some marketing uses:

• Social media content: Perfect for Instagram Reels, TikTok, and YouTube Shorts

• Product reveals: Turn static product shots into dynamic showcases

• Brand storytelling: Create mood pieces and atmospheric content

• Campaign concept testing: Rapidly prototype video concepts before investing in full production

In terms of limitations, videos are generated in 480p quality (standard definition), and the aspect ratio depends on your starting image. You're also limited to relatively short clips. 

For social media and digital advertising, aesthetic quality often matters more than technical specifications – so get experimenting with product images, brand assets, etc. 

Honestly, some examples are crazy. Once these models can output 1min+ of footage…wow that’s really going to challenge a lot of industries. 

AI agents become more usable

The AI agent space has been full of grand promises and disappointing deliveries. There’s three interesting projects of late:

MiniMax agent

MiniMax Agent enables you to create custom AI agents using only natural language descriptions. You define what you need the agent to accomplish. Some key uses:

• Customer support chatbots: Create agents that handle FAQs, route tickets, and manage onboarding processes

• Content creation assistants: Build agents that write blog posts, social media content, and product descriptions tailored to your brand voice

• Research agents: Deploy agents that compile industry data, analyze competitor activities, and track market trends

• Lead qualification systems: Create agents that screen prospects, score leads, and route them to appropriate sales team members

MiniMax’s below thread offers some excellent insights and uses.

Manus

Unlike traditional chatbots that wait for instructions, Manus operates as a truly autonomous agent. You set a goal, and it breaks down the task, navigates the web, gathers information, and delivers complete outputs, such as spreadsheets, reports, and analyses.

The transparency element is cool. The "Manus's Computer" window allows you to observe the AI's decision-making process in real-time and intervene when necessary.

MIT of all institutions tested it (no pressure):

• Research compilation: Successfully created a comprehensive list of 30 tech journalists with current employers and notable work

• Property search: Found two-bedroom listings in NYC with relevant details and market context

• Academic nominations: Struggled with complex research across paywalled content but eventually delivered solid results

Key quote: "I was impressed that I was able to make top-level suggestions for changes, much as someone would with a real-life intern or assistant, and that it responded appropriately."

They’ve just released their “Cloud Browser” for carrying out autonomous tasks on the internet.

Proactor AI

Proactor doesn’t wait for you to ask questions or give commands. It proactively participates in your conversations and meetings.

It can join your Zoom calls, Google Meets, or in-person conversations and provides real-time transcription, identifies actionable tasks, and delivers insights automatically.

Some key use cases:

• Sales teams: Real-time call coaching, BANT qualification, objection handling suggestions

• Recruiting: Live candidate evaluation, question suggestions, skill assessment against job requirements

• Project management: Automatic task identification, timeline tracking, decision documentation

• Content creation: Research suggestions, fact-checking, story angle development

Proactor is still in waitlist mode with no public pricing announced. But early user feedback is looking good, with many calling it the first AI tool that feels like a genuine collaborator. 

Phew! Pretty dense one this edition! But hopefully something for everyone.

AI MARKETING EVENTS

🗓️ Recommended AI marketing events

Whether you’re building with AI, leading a marketing team, or just trying to stay sharp in a fast-moving space, events are still one of the best ways to plug in.

Great events do more than just deliver content — they give you context. You get to see what other teams are testing, what’s actually working, and where the market is headed.

They’re also one of the few places where you can meet like-minded people, talk shop with folks solving the same problems, and build a network that isn’t purely algorithmic!

Some top events to check out:

•  AI Agents Summit (Virtual) — September 18–19, 2025: Laser-focused on AI agents, copilots, and autonomous systems. If you're experimenting with automating parts of your marketing workflow or interested in agent-based design, this one’s worth attending.

• AI for Marketers Summit (Virtual) — November 13–14, 2025: Created specifically for marketing professionals. From prompt engineering to campaign automation, it’s a practical look at how AI is being used right now in real teams.

• Data + AI Summit (San Francisco) — June 9–12, 2025: Hosted by Databricks, there’s plenty of marketing relevance here — especially around LLM workflows, AI tooling, and future-proofing your stack.

• Ai4 2025 (Las Vegas) — August 11–13, 2025: A big-picture AI conference covering finance, healthcare, retail, marketing, and more.

Hope you enjoyed this week’s issue. If you missed the last newsletter, you can read it here.

If you found it useful, please recommend it to a friend or colleague.

Until next time. Happy marketing.

—The AI Marketer