How Kalshi’s wild NBA Finals ad signals a new era of fast, chaotic, AI-powered commercial production, and what it means for creative pros.
First came Coca-Cola’s glossy AI nostalgia trip. Then came Kalshi’s NBA Finals ad, which was put together in just three days, for less than the cost of a new iPhone. Fast, chaotic, and unmistakably different, it didn’t look like a traditional commercial. And that was the point.
Welcome to the new era of ad production. With tools like Envato VideoGen, AI-generated ads are changing the way creatives work. If you’re a motion designer or video editor, welcome to your new creative playground.
From polished to surreal: How AI-generated ads are evolving
Coca-Cola’s AI-powered ad for Christmas 2024 was sentimental and clearly backed by a big budget. The spot leaned into familiar branding, with lifelike graphics and carefully orchestrated scenes. It was a safe, high-production-value entry into the AI space, and for many, it marked the moment major brands started taking AI seriously.
Then came Kalshi’s NBA ad. Chaotic, glitchy, and kind of unsettling, it looked nothing like Coke’s. Made in 72 hours for just $2,000, it felt more like a digital fever dream than a traditional commercial. But it got people watching, sharing, and searching.
Kalshi’s NBA ad and the rise of scrappy creativity
“Kalshi hired me to make the most unhinged NBA Finals commercial possible… Network TV actually approved this GTA‑style madness.” — PJ Accetturo, Director of the Kalshi ad (Business Insider)
Built almost entirely with AI tools (including generative visuals, text prompts, and synthetic voiceover), the Kalshi ad demonstrated that AI commercial production doesn’t need polish to punch through the noise.
Instead of smoothing over AI’s rough edges, Kalshi leaned into them. The result felt like a weird digital hallucination, and that was precisely what they were going for!
Low cost, high impact
“The client got an insane ad for a great rate on a blistering timeline, and I got paid really well, while working in my underwear.” — PJ Accetturo, Director of the Kalshi ad (Business Insider)
Kalshi’s tiny $2,000 budget didn’t hold it back. If anything, it gave the team creative freedom. No cast, no crew, no physical sets: just a small group of creatives, a few prompts, and AI tools doing the heavy lifting. The result was wild and unpredictable, perfect for standing out in a sea of polished ads.
Shoot anywhere (without leaving your desk)
One of the ad’s most striking qualities was its world-building. It jumped between hyper-surreal sports courts, people in egg baths, and glitched-out jungles in just a few seconds. These scenes would have been expensive and even impossible with traditional production methods. With AI tools, the Kalshi team could create whatever they wanted, instantly:
- No location scouting
- No permits or travel
- No weather delays or reshoots
- Complete creative control over space, lighting, and mood
Fast iterations, fast turnaround
The most significant innovation was in the workflow. Kalshi’s team built the ad using fast-turnaround, generative-first production cycles to quickly test ideas and create variations. The iterative process was about letting the weirdness evolve, and then curating the best moments into a tight, shareable spot.
“This took about 300–400 generations to get 15 usable clips… One person, 2‑3 days. That’s a 95% cost reduction vs traditional ads.” — PJ Accetturo, Director of the Kalshi ad (The Verge)
What visual chaos really means and why it works
A new visual language is taking shape in the new world of AI-generated ads. It feels fast, surreal, and sometimes a little unhinged. Forget polish or perfection: this emerging aesthetic, often described as “visual chaos,” is about grabbing attention in a split second. It usually looks something like this:
- Glitchy transitions
- Surreal, disjointed visuals
- Emotionally jarring cuts
- Uncanny valley vibes
Many are calling this visual chaos, and it’s quickly becoming the signature aesthetic of AI-generated ads. However, the truth is, it’s not all artistic direction. Much of the chaos is built into how these AI tools function.
Why visual chaos is (currently) baked in
At this stage of generative video evolution, visual chaos is more a byproduct of limitations than a fully intentional design language. Tools like Envato VideoGen are making massive leaps, powered by a lineup of AI models each tailored to different kinds of visual outputs — but they’re not magic.
- Clip length limitations: Most AI video tools generate short, isolated clips (often under 10 seconds). Stitching them together into a coherent sequence creates jumpy and jarring transitions.
- Lack of narrative memory: AI can’t yet maintain consistent characters, motion, or perspective across shots unless heavily engineered.
- Rendering constraints: The more coherent and controlled you want a shot to be, the longer it takes to generate.
- Post-production time: When building a commercial in 72 hours (like Kalshi), you don’t have time for smoothing, masking, or VFX cleanup. So the quirks stay in.
In other words, the chaos is necessary when you’re moving fast with today’s tools.
How speed affects visual chaos
When working with a tight turnaround (like Kalshi’s 72-hour window), every decision is a trade-off between control and momentum. Under these conditions:
- You’re relying on first or second-pass generations: AI tools often require several prompt iterations to get something usable. But when time is short, there’s no luxury to wait for the “perfect” output. You go with what works or what works well enough.
- There’s little time to finesse transitions or ensure consistency: In traditional post-production, editors can mask inconsistencies, refine pacing, and match visual styles from shot to shot. With AI output, mainly when clips are generated independently, that level of polish often isn’t feasible under pressure.
- Re-rendering takes time: Envato VideoGen can produce excellent results, but regenerating footage isn’t instant, especially when the prompt changes mid-way through a concept. Iteration slows down when generation times climb.
- It’s more efficient to lean into the chaos than fight it: Instead of trying to smooth out every visual inconsistency, many creators now treat those quirks as part of the creative direction. Glitches become visual metaphors. Discontinuity becomes a style.
This is precisely what made Kalshi’s approach stand out. The ad’s creators didn’t try to hide the imperfections or force traditional narrative logic. Instead, they embraced the unpredictability of the medium, turning what could have been limitations into a deliberate, scroll-stopping aesthetic.
How visual chaos might evolve from here
As generative video technology matures, the visual chaos we see today may fade into the background. The current aesthetic (marked by abrupt transitions, uncanny visuals, and surreal pacing) is shaped by technical limits, but those limits are already being pushed.
The AI tools of the future will generate longer, more coherent scenes with smoother motion and better character consistency, reducing the need to stitch together disjointed clips. When that happens, visual chaos will become a choice, not a constraint.
How to get stuck in with your own AI-powered ad
So, how do you actually create something like Kalshi’s NBA ad? Tools like Envato VideoGen make the process fast, flexible, and surprisingly fun. Here’s how to approach things:
Step 1: Access VideoGen
First, head to Envato VideoGen and sign in with your paid and active Envato account. You’ll land on a clean, simple welcome screen where you can start a new video.
Step 2: Choose your aspect ratio and audio
Select the aspect ratio you want for your final video, and ensure that the Audio toggle is switched on if you want your final video to have sound (note that only 16:9 is available with audio).
Step 3: Write your prompt
Here’s where your prompt comes in. Add a detailed description of the scene you want to create in the prompt box. Your prompt is your creative foundation. The stronger it is, the better the AI output.
- Be specific: “a flooded city at night lit by neon” will outperform “urban scene.”
- Add emotion or tone: “dreamlike,” “apocalyptic,” or “satirical” can steer the vibe.
- Use unusual combinations: to trigger fresh results (e.g., “sci-fi opera set in a courtroom”).
Step 4: Generate visual sequences
Let the AI generate a variety of different clips by using different prompts.
- VideoGen can quickly generate multiple versions, so make use of this.
- Don’t expect perfection — look for standout moments or surprising imagery.
- Embrace randomness. The unexpected often becomes the most compelling part of your story.
Step 5: Edit and refine
Once you have your clips, gather them and use a video editor like Adobe Premiere Pro to start shaping your narrative.
- Trim and arrange the clips to create flow and contrast.
- Focus on rhythm and emotion rather than linear storytelling.
- Use intentional dissonance, jumps, glitches, or tonal shifts to hold attention.
- Layer in transitions and overlays from Envato to enhance the look.
Step 6: Add audio and branding
Add sound and identity to bring everything together.
- Use AI voiceovers or royalty-free tracks from Envato to set the tone.
- Add sound effects or music to match the energy of your visuals.
- Apply branding elements: logo animations, type treatments, and brand colors.
- Keep it cohesive, but don’t be afraid to clash styles if it adds impact.
Why big brands are getting involved, and where they’re holding back
Coca-Cola showed that AI could be polished. Kalshi proved it could be disruptive. And together, they’re part of a growing shift in how brands can approach commercial storytelling.
AI is lowering the barrier to experimentation
For decades, ad production has been a high-stakes, high-cost process. This is especially true for prominent brands with strict identity guidelines and international visibility. But generative AI is shifting that dynamic.
Lower production costs mean brands can run experimental campaigns without risking multi-million-dollar budgets. Meanwhile, rapid iteration allows them to A/B test messaging, visuals, and tone in real time, which is increasingly valuable across fast-moving platforms like TikTok and Instagram.
AI tools also make hyper-localisation at scale possible. Brands can now generate dozens of region-specific ad variations without filming multiple versions.
Why polish still matters to big brands
While smaller players like Kalshi are willing to embrace glitchy aesthetics, bigger brands are still hedging their bets. Most don’t want to risk damaging brand trust with ads that feel too unpolished or uncanny, but some have faith in the process. Unilever, one of the world’s broadest multinationals, recently announced the launch of over 20 AI-driven studios, charged solely with creating marketing content for its homecare brands.
“We want these teams to truly understand our brands end-to-end, because the quality of the output will be better. We’re trying to elevate the design capabilities.” — Mario Dughi, Unilever Global Marketing Director (Design Week)
AI’s current visual instability may create trust issues if not carefully managed. This is why big brands like Coca-Cola and Unilever are only dipping their toes into AI through tightly controlled, hybrid workflows where generative content is either used in the concept phase or heavily refined before going public.
Here’s why polish still matters:
- Visual consistency builds trust, especially in markets like finance and health.
- Brand guidelines aren’t optional, and AI still struggles to reliably follow layout, typography, and tone without heavy prompting and oversight.
- Diversity and representation require intentionality, and without careful direction, AI outputs can default to biased or homogenized imagery. Maintaining inclusive, thoughtful visuals requires human control.
“First, it’s not like you input one prompt saying ‘create an ad’ and it pops out an ad. A lot of creative decisions are all made by humans.” — Pratik Thakar, Global Head of Creative Strategy, Coca-Cola (Marketing Dive)
Where big brands are headed next with AI
While most big brands are still using AI cautiously (testing it in concept work, social content, or experimental campaigns), they’re also laying the groundwork for deeper integration. Here’s how it could potentially look for big brands using AI in the next few years.
- Custom-trained brand models: To maintain control, many companies are now exploring fine-tuned AI models trained on their brand guidelines, tone of voice, and visual identity. These models could generate on-brand content more reliably than general-purpose tools.
- AI-powered internal content studios: More brands could potentially bring generative tools in-house. With the right talent and safeguards, internal teams can produce branded content faster and cheaper, without sacrificing quality.
- Bringing in AI during development: Some bands will begin using AI in the early creative strategy process (concept exploration, moodboarding, ideation, and even data-informed audience targeting).
“More commonly, many marketing departments and their agencies are now using generative AI to quickly spin up hundreds of different iterations of ads in a fraction of the time it would take for a human to produce them” — (Business Insider)
How motion designers can ride the chaos
If you’re a creative pro, AI is a potential turbo boost. Use it as your collaborator, and focus on the stuff machines can’t do: narrative, emotion, rhythm. Let the bots handle the grind; you shape the spark.
Here’s what AI can take off your plate:
- Generating visual concepts from prompts in seconds
- Filling in missing shots or transitions during the edit
- Producing multiple variations for rapid iteration and testing
Keep control, and move faster
Tools like Envato VideoGen give you the power to move quickly without sacrificing creative direction. Start small, experiment often, and follow the unexpected. The visual chaos is part of the process, but you’re still calling the shots.
This is creative evolution on fast-forward
AI is changing how ads get made and what they look like. The past year has shown how fast we’re moving, from Coca-Cola’s careful CG to Kalshi’s chaotic slam dunk.
The takeaway? Don’t wait for perfection. Embrace the weird, move fast, and start creating. This is the new creative playground!
FAQ
What is an AI-generated ad, and how is it different from traditional ads?
AI-generated ads are created using tools that automate or assist with visual generation, editing, and scripting. They’re often faster, cheaper, and more surreal than traditional productions.
What’s the difference between the Coca-Cola and Kalshi AI ads?
Coca-Cola’s ad was polished and cinematic. Kalshi’s was chaotic and experimental. The gap shows how quickly AI ad styles are evolving.
What does visual chaos mean in an advertising context?
It deliberately uses surreal, unpredictable visuals to grab attention—think glitch art meets ad copy.
How can I use Envato VideoGen to make an AI ad?
Start with a strong prompt, generate visuals, edit the sequence, and add music/branding using Envato VideoGen and Envato assets.
Will AI replace motion designers or help them?
It’ll help them. AI can handle repetitive tasks, giving creatives more time to focus on storytelling and innovation.
