The Taste Gap

Why creative edge (not AI) will define the next generation of standout brands

AuthorAuthor

Peter Benei & Torsten Sandor
June 17, 2025

For years, I’ve told people I live in Tuscany not because it’s my home country, but because this corner of the world has taste.

When the entire world will be flooded with tasteless autogenerated garbage content, this place, this only, will be the last domino standing. I’m 100% sure of it.

Taste. Not the showy kind. Not that “modern” madness. The kind that’s embedded in everything. You feel it in the food. In the quiet richness of the Renaissance. In the legacy of the Medici, the works of Da Vinci, the wines of Chianti, and the stillness of the Versilia coastline, where I live now. I chose this place because I wanted to be surrounded by taste. And now, finally, I can connect that personal preference to my professional obsession.

Taste is the ultimate edge. The last frontier.

In the age of AI, everyone has the same tools. Same models. Same prompts. But not everyone has taste. Taste is the final advantage. The last real moat. It’s the difference between a spot that breaks the internet, and one that looks like a Fiverr job. Between a brand that lingers, and one that scrolls by, unnoticed.

This week’s edition is a love letter to that edge. AI is flooding the creative zone. But the winners? They’ll be the ones who know what good looks like, and who build systems to scale it.

Let’s get into it.

Peter Benei | Co-author

Consultant at Anywhere Consulting

Torsten Sandor | Co-author

Senior Director of Marketing at Appen

A quick word from this week’s sponsor:

Turn AI into Your Income Engine

Ready to transform artificial intelligence from a buzzword into your personal revenue generator

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

  • A curated collection of 200+ profitable opportunities spanning content creation, e-commerce, gaming, and emerging digital markets—each vetted for real-world potential

  • Step-by-step implementation guides designed for beginners, making AI accessible regardless of your technical background

  • Cutting-edge strategies aligned with current market trends, ensuring your ventures stay ahead of the curve

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

AI Takes the NBA Finals

Prime-time video. Zero production crew.

Prediction-market platform Kalshi didn’t just buy a 30-second ad slot during Game 3 of the NBA Finals. They redefined how it gets filled. They hired PJ Accetturo (aka PJ Ace) to create the entire spot solo, powered by AI.

No agency. No studio. Just a laptop, Gemini, and Google Veo 3.

In 48 hours, he:

  • Drafted 10 offbeat character vignettes with ChatGPT/Gemini

  • Generated 400+ clips using Veo 3

Edited and delivered a finished spot — all for ~$2K

He didn’t stop there. Accetturo open-sourced the entire process: from prompt strategies to color-grading hacks. And he left CMOs with a mic-drop takeaway:

❝

Just because this was cheap doesn’t mean anyone can do it… Brands still pay a premium for taste.

Meanwhile, ByteDance launched its Seedance 1.0 model — already beating Veo 3 on the smoothness and shot consistency. It costs ~$0.50 per 5-second render...

And in research mode, they unveiled real-time AI video generation via Seaweed-APT 2. Yes — video, streamed live, generated on the fly.

Footage from the Kalshi video

Why it matters for marketers

This is the future of branded video — faster, cheaper, and more reactive than ever.

But it’s not plug-and-play.

The new model assumes:

  • You have a director’s eye or someone who does

  • You’re ready to publish weekly, not quarterly

  • You’ve updated brand rules for synthetic content

Without that, you’ll miss the upside — or worse, ship off-brand garbage.

The next wave of content won’t come from agencies.

It’ll come from internal micro-studios: one creative, one prompt engineer, one editor — publishing weekly, riding cultural spikes, and building brand equity at startup speed.

What to do next?

Run a $10K ad sprint
: Task your brand lead with remaking your 2024 hero spot using Veo or Seedance. Compare impact and recall.

Pre-build your prompt library: 
Create ready-to-use templates for launches, promos, and cultural moments so your team can move on a day’s notice.

Update your governance: Get your legal and brand teams to weigh in on likeness rights, copyright use, and SAG/AFTRA compliance — before someone ships an all-AI cut.

Anthropic’s Blog Bombs

The Lesson Is Clear

Last week, Anthropic — maker of Claude — launched a blog written entirely by AI.

This week? It’s gone.

The company quietly pulled the plug after backlash labeled it “soulless,” “robotic,” and the worst offense of all: boring.

Critics called it the poster child of automated content marketing gone wrong.

Anthropic “Claude Explains” since then dead blog

Why it matters for marketers

This is the clearest signal yet: AI can write. But it can’t connect.

Anthropic’s misstep proves what your audience already knows — they can spot lifeless, inauthentic content a mile away. And they’ll tune it out instantly.

Automation can support content. But only humans can infuse it with insight, relevance, and brand voice.

In a world where everyone has AI, authenticity is your moat.

What to do next?

Define where AI belongs. Don’t ban it — just draw clear lines. Use AI for briefs, outlines, and data pulls. But never let it hit “publish” without a human pass.

Codify your voice. If your brand doesn’t have a tone-of-voice doc yet, create one. It’s the creative constitution your team (and your AI) should follow.

Empower your editors. Writers become directors. Their job now is curating, refining, and humanizing. Invest in that skillset — it’s the last mile that makes content resonate.

OpenAI’s New Strategist Model Just Dropped

Think less chatbot, more chief strategy officer.

OpenAI has launched GPT-4o-pro (aka “o3-pro”), its most advanced AI model to date — now available for ChatGPT Pro and Team users.

This isn’t about faster answers. It’s about better ones. o3-pro literally “thinks before it speaks,” taking time to run deep reasoning before delivering responses.

In OpenAI’s own words, it beats Google Gemini 2.5 Pro and Anthropic Claude 4 Opus on advanced strategic tasks. On launch day, OpenAI also slashed the cost of the standard o3 model by 80% — making it the new default for high-volume AI work.

How we’re using it

We uploaded a CSV of Meta ad spending from this quarter and asked o3-pro to find waste, wins, and optimizations.

It gave us:

  • A 7-day test matrix An early-warning creative fatigue dashboard

  • A channel-level ROI breakdown with tactical next steps

Here’s a (lightly redacted) snippet from its reply. We are impressed.

Our use case output from o3-pro

Why it matters for marketers

This model isn’t just chat — it’s strategy.

For $200/month, you now get access to a tool capable of:

  • Market mapping

  • Competitive intelligence

  • Scenario modeling

  • Strategic planning

Stuff that previously required high-priced consultants or your most senior team members.

But it’s only as good as the context you feed it. Garbage in, garbage out.

What to do next?

Learn how to prompt it: Use this quick-start guide to understand o3-pro’s expectations. Share it with your team. This model thrives on detailed input.

Test it on a real challenge. Give your strategy team a live prompt: find a new market, war-game a competitor’s move, or optimize an underperforming funnel.

Audit your consultant spending. Review what you’re outsourcing for research and planning. If it’s pattern-based or data-heavy, o3-pro may already be good enough — in-house.

Mattel x OpenAI

AI Becomes the Product

Mattel just partnered with OpenAI — and it’s not just about operations. The toy giant plans to embed generative AI into the heart of its business: toy design, interactive content, and new forms of play across brands like Barbie and Hot Wheels.

Think less "automating workflows" and more "reimagining the product."

Why it matters for marketers

This is your roadmap for next-gen AI adoption. Mattel isn’t just using AI to market toys — they’re using it to make toys smarter, more personalized, and more interactive.

That’s the unlock: AI as a product experience, not just a productivity boost. The most innovative brands will build AI into what they sell — not just how they sell it.

What to do next?

Bring product and marketing together. Run a joint session with product, R&D, and GTM. Ask: “What would it look like if our product could learn, adapt, or interact in real-time?”

Prototype something small. You don’t need a giant budget. Start with a smart quiz, AI-powered explainer, or interactive buyer guide. The key is testing how intelligence changes the experience.

Look beyond your category. Watch how consumer, retail, and gaming brands are embedding AI. The biggest insights often come from industries one step ahead.

ElevenLabs v3

Your Brand Can Now Whisper, Laugh, and Sigh

ElevenLabs just dropped v3 — and it’s more than an upgrade. It’s a leap into emotional AI audio. The alpha version introduces:

  • 70+ languages Multi-voice dialogue

  • Inline tags for emotional tone — like [whisper], [laughs], [sighs]

Think podcast-grade delivery — but fully synthetic.

Yes, it demands more prompt finesse than earlier versions. But the reward? Film-quality nuance, ready for global scale.

Why it matters for marketers

Voice is the fastest way to scale across borders. Now a single product script can ship in Portuguese, Arabic, and Mandarin — complete with tone and emotion — before your design team wraps the thumbnail.

Pair this with AI video, and you've got a global, studio-grade campaign with zero studio time.

What to do next?

Rethink customer service. Use dynamic voice for IVR flows that feel more human. "[Sighs] I can see that’s frustrating" lands better than a robotic apology.

Upgrade sales content. Create regional voiceovers for case studies and product intros — localized in accent, tone, and language.

Launch internal “memo-casts.” Have your execs drop weekly audio briefings — voiced by their AI twins, reviewed for tone. Great for alignment, even better for leadership presence.

Inside the Perplexity Fellowship

How to build AI ops into your organization

As part of the Perplexity Business Fellowship Program, we joined a fireside chat with Aaron Linsky, CTO of AIA Labs at Bridgewater.

Here are our 8 takeaways from the workshop — and what marketers should be doing now:

1. Don’t just add AI to existing teams
Disposable credit cards and sandbox testing: new tools are tested in a secure environment without real data. If a tool shows early signal, it moves to real use cases.

» What marketers can learn:
Spin up your own AI “lab” with zero-risk test zones. Let curiosity, not bureaucracy, drive tool discovery.

2. No hype, just use-case proof
They only keep tools that solve real business needs. “Would people be upset if we removed it?” is a legit metric.

» What marketers can learn:
Don’t waste time debating AI tools. Build something useful, and the adoption will prove itself.

3. APIs > UIs
They skip glossy apps and build with APIs for control, speed, and custom workflows.

»  What marketers can learn:
Use APIs to embed AI into daily ops. Don’t wait for enterprise tools to catch up to your needs.

4. Three data types fuel AI performance
Public web (e.g., Perplexity), internal docs (e.g., SharePoint), and third-party intelligence (e.g., Bloomberg).

» What marketers can learn:
Audit your data stack. AI impact = the quality + diversity of your unstructured data inputs.

5. AI-powered deep research converts skeptics
The killer use case: high-quality, cited answers to complex questions in minutes instead of hours.

» What marketers can learn:
AI shouldn’t just summarize. Use it to deliver strategic insights and decision-ready answers.

6. Chart-reading and dashboard building is automated
LLMs parse screenshots and generate dashboards (HTML/Plotly) from scratch.

» What marketers can learn:
Build AI agents that turn messy visuals into decision tools. Think: one-click exec summaries.

7. “Vibe coding” makes builders out of anyone
Tools like V0 and Replit let non-devs ship working apps with plain English.

» What marketers can learn:
Empower your team to prototype fast. Internal tools no longer need a dev backlog.

8. Don’t convince skeptics. Win with results.
Early adopters get support, not resistance. Success stories turn critics into champions.

» What marketers can learn:
Focus on pilot wins. Adoption spreads fastest when people feel the productivity leap.

Need access to Perplexity Pro for free?

As business fellows, we have a few invites for Perplexity Pro.

Rules are as follows:

  • You will be added to our consulting organization’s account

  • You have Pro access until the end of the year

  • You can’t upload files, documents, use APIs, or invite others - these are the restrictions due to usage limits

Aside from these limits, invites get you a free Perplexity Pro account, no need to pay anything at all.

If you are interested, simply reply to this email.

Rapid-Fire News

AI News You Need to Know

What Just Happened in AI (That You Should Actually Care About)

1. Anthropic Drops Free AI Fluency Course
Anthropic just launched a no-cost “AI Fluency” course. Ideal for leveling up your entire org, fast.
» For marketers: Make this mandatory onboarding. AI literacy across departments is now table stakes.

2. AI Is Fueling the Next Wave of Tech Layoffs
U.S. tech unemployment is climbing, with AI-driven restructuring as the likely culprit.
» For marketers: Shift from headcount to leverage. Start reskilling your team for AI-native roles before someone else decides for you.

3. Disney & Universal Go After MidJourney
A new copyright lawsuit targets AI image generation. This time with Hollywood firepower.
» For marketers: Review your AI media policy today. The legal risk around visuals is no longer hypothetical.

4. Mistral’s New Models Are Built for Europe
French AI leader Mistral released two new reasoning models optimized for multilingual tasks.
» For marketers: Global brand? Use Mistral for campaigns and research in European languages. It may outperform U.S.-centric models.

We’ve officially hit the point where your marketing team can launch a global campaign, run deep strategic analysis, and ship studio-grade video. All from a laptop.

But tools alone won’t differentiate you. Taste, speed, and execution will.

AI gives you leverage. Your job is to turn it into momentum.

» Forward this to someone still building campaigns like it’s 2019.
» Ready to rethink your stack, strategy, or team? We help B2B companies do exactly that.

Until next week,
Peter and Torsten