Can AI Crack the Code of Cannes Lions? Our 2025 Titanium Predictions
We have the world's most comprehensive, independent advertising database (4,400 award-winning ads from the last 7 years). So we thought we should probably do something interesting with it, right?
Cannes Lions is undoubtedly the most prestigious award show for the creative industry around the world. It's where careers enter two modes: "make" or "break."
So last year we decided to embark on this impossible journey into the unknown: can AI crack the winning formula?
Not just from how submissions are crafted, but from the intangible insights you can only learn from real jurors who've sat in those rooms, felt the energy and made the calls.
It's our most challenging experiment of AI limits in advertising, yet. And that's exactly why we had to try.
The 2024 Experiment: Better Than Random
Last year, we focused on Titanium Lions and built our first prediction algorithm.
Out of 4 winners, we got 2 right:
✅ DoorDash "All-The-Ads" – Predicted Titanium, won Grand Prix
✅ JCDecaux "Meet Marina Prieto" – Predicted Titanium, won Titanium
❌ DP World "The Move to -15" – Missed it
❌ Xbox "The Everyday Tactician" – Missed it
50% accuracy. Not bad for predicting human creativity, but we wanted to see if we could do better.
2025: Building a Smarter System
Instead of just analyzing Titanium winners, we studied every Gold and Grand Prix campaign from 2024 to understand what actually wins at Cannes.
We had an advantage: detailed descriptions of all 400 Cannes 2024 winners in our database, including the 99 Gold and Grand Prix campaigns across all 32 categories. This made building a more sophisticated scoring system surprisingly straightforward.
The complexity jump was massive: analyzing 99 winners from 32 categories judged by 290 jurors from 53 markets vs just one small category judged by 11 jurors. That's a 900% increase in data complexity.
As someone who's always kept Lions close to my heart – maybe not obvious for a strategist, but I even won a couple while running my own agency – I've been obsessed with thought experiments about what AI can and cannot do in advertising.
Predicting Cannes is by far my favorite, while being the most futile.
From this analysis, we developed the "Cannes Caliber" scoring system.
The Cannes Caliber Framework
Our algorithm scores campaigns across four dimensions:
Transformative Impact (40 points) - Market disruption and cultural influence
Groundbreaking Creativity (30 points) - Originality, innovation, and craft
Strategic & Cultural Relevance (15 points) - Brand alignment and audience resonance
Compelling Communication & Emotional Impact (15 points) - Clear communication and emotional connection
Here's what made this different: Our AI didn't just score campaigns – it developed "Cannes Caliber" as a special distinction for work that truly pushes the industry forward, beyond just being good advertising.
Our Method
We used Gemini 2.5 Pro to analyze the winners using our scoring framework.
We created a custom GPT trained as a Cannes Lions juror with 15+ years of industry experience, which we've made available to everyone here (it offers simplified scoring for winners plus suggestions based on 2024 Lions winners).
Each shortlisted campaign was analyzed 10 times, we discarded the highest and lowest scores, and averaged the rest. Campaigns scoring 90+ earned our Titanium prediction.
Full results table:
Our 2025 Titanium Predictions
1. Caption With Intention (91.3 points)
Academy of Motion Picture Arts & Sciences | FCB Chicago
Problem: 430+ million Deaf/HoH people experience incomplete cinema due to static captions unchanged since 1971.
Idea: Dynamic captions with expressive typography, real-time sync, and character color-coding.
Why our AI thinks it wins:
This transforms captions from utility into art form. By making typography express emotion and sync word-by-word, it creates a new accessible visual language for cinema. The Academy's integration into Oscars rules means this isn't just innovation — it's institutional transformation that redefines accessibility for 430 million people globally. Early pilots show a 94% jump in scene comprehension among Deaf viewers.
2. The Flame that Wasn't a Flame (91.0 points)
EDF | BETC Paris
Problem: The Olympic flame symbolizes tradition but conflicts with sustainability commitments.
Idea: 100% electric "flame" using 200 misting nozzles and 4 million LED lumens.
Why our AI thinks it wins:
Pure conceptual audacity: reimagining the Olympic flame — unchanged for decades — as the world's most visible proof that sustainability doesn't require sacrifice. Using 200 misting nozzles and 4 million LEDs created something more beautiful than fire while positioning EDF's electric future at the emotional heart of the Games. Turns the world's most sacred sporting symbol into a climate-action billboard, driving 15% lift in EDF brand-preference across France.
3. Caramelo (91.0 points)
Pedigree | AlmapBBDO São Paulo
Problem: Brazil's beloved street dogs have 90% lower adoption rates than purebreds
Idea: Give Caramelos an official 'pedigree' with DNA research, certificates, and kennel club
Why our AI thinks it wins:
Brilliant inversion of the "pedigree" concept itself. Instead of accepting that mixed breeds lack status, Pedigree created official legitimacy through DNA research and institutional recognition. The 218% adoption increase and 690M social impressions prove it worked, but the real genius is using brand equity to rewrite social hierarchy — turning underdogs into the most desired dogs in Brazil. Execution spanned DNA testing, official certificates, Kennel Club partnership, and custom packaging.
4. Three Words (90.0 points)
AXA | Publicis Conseil Paris
Problem: In France, a woman dies every 3 days from domestic violence, 89% at home.
Idea: Add "and domestic violence" to home insurance policies as covered emergency.
Why our AI thinks it wins:
Every three days a woman in France is killed by domestic violence — usually at home. AXA embedded this reality into the legal framework of home insurance, treating it with the same institutional seriousness as natural disasters. Three words — "and domestic violence" — updated 2.5 million policies and helped 121 women — in month one. It's business model innovation disguised as copywriting, with France's insurers' association now drafting national guidelines.
What We Actually Uncovered
This difference between industry impression and juror scrutiny became clear with Penny's "Price Packs" – highly praised in media coverage for its clever "price as design" concept, but scoring just 71.9 in our system.
When you watch case study videos, it's easy to be swayed by smart ideas. But Cannes jurors apply rigorous analysis, looking for transformative impact and cultural resonance beyond just clever execution. Our AI tries to simulate that deeper scrutiny, not just first impressions.
Combined with our understanding of the impossible-to-model human factors – jury dynamics, presentation energy, the intangible "X-factor," who speaks up for which work during deliberations – we learned that even perfect data isn't enough to predict creative excellence.
But it can help us understand the patterns behind it.
Expert Panel
“I like to think of your Cannes-cracking bot as more of a diagnostic tool than just a predictive one: perfect for spotting where every other creative is headed so you can maybe get one step ahead.
I couldn't not bring up Goodhart, because it would be crackers to not heed the warning that those neat score-clusters mark yesterday’s flavours - extremely useful signposts, but not ideal destinations.
Treat the patterns as a treasure map to finding the gaps, pointing to the blank squares no jury can even imagine yet. Maybe AI can draw the grooves so we can veer gleefully off it with something award-breakingly new.”
Joe Burns, Strategy Lead at Quality Meats Creative
“The premise that creative merit is objectively quantifiable is a shared delusion we all agree to because awards feel good and agency decks need logos. But let’s not pretend: every jury is a human algorithm of aesthetic bias, cultural baggage, strategic compromise, and subconscious envy. There is no Platonic ideal of a “deserving ad”—there’s just a group of people in a hotel ballroom in Cannes, trying to decide if a metaverse activation for oat milk is cooler than a PSA that made them cry once in March.
Which is why Adaily’s predictions are weirdly necessary. They won’t be accurate in the way stock tips or weather apps aim to be accurate—but they might accidentally reveal more about the juries than the work itself. And maybe, in the margins of all that educated guessing, they’ll accidentally spotlight a piece of creative ephemera that truly deserved better. Or maybe not. But that’s the game, isn’t it?”Jake Neske, Managing Director, Hustle and Also Known As
“This is seriously impressive work. Easily the most comprehensive, data-driven attempt I’ve seen at decoding what makes Cannes-winning work. The methodology is rigorous, the scoring framework is smartly structured, and the predictions feel grounded in real jury logic.
I’d love to see this level of detail extended to all the other Cannes Lions categories, where emotion, craft, or strategy play different roles.
That said, Cannes will always be more than data. The jury room is where the real magic happens: the late-night debates, the passionate defenses, the gut feelings, and sometimes even the politics. No algorithm can fully replicate that emotional, human reception. At least not yet. But if one day an AI ever does hand out awards, it should absolutely be powered by Adaily.”Matthieu Lamoureux, Editor of LLLLITL and Strategy Director at Jellyfish Amsterdam
“The novel approach of creating a Cannes Caliber Framework is fascinating! So often we wonder why certain ads won at Cannes, but being able to quantify some of the intangible things we know the jurors care about is so smart. In particular the custom GPT will make it easier for creatives to judge their own work in preparation for a submission to Cannes as they’ll be able to see what changes might need to be made to their case study video. Excited to see how it all plays out!"
Jeff MacDonald, Director of AI at Mekanism
“It may be novel to use AI to improve ad creative now, but it will be commonplace in the not so distant future. But AI shouldn’t be the final decision maker. It should be a partner that helps human creators to achieve new levels of creativity and ad effectiveness.”
Debra Aho Williamson, Founder and Chief Analyst at Sonata Insights
“What Piotr has done here is really smart, because it's not about prediction (although those are fun to geek out over). It's about AI's ability to surface structural patterns and strategic insights that would take humans ages (if at all). Think of it like an audio equaliser, AI is not about amplifying everything across the board, but boosting specific frequencies that matter: transformative impact, cultural resonance, systemic change. The kind of strategic intelligence that can inform better creative decision-making from day one, minus the scary bits about machines taking over. It's AI as creative co-pilot, not creative replacement.”
Andrew McLean, Founder at garden gnome
The Real Experiment
Even though we can collect all the data on previous winners, we know it's not enough to predict the future. This is more of a thought experiment about how well AI can simulate juror behaviors, weighing measurable elements against intangible factors that exist outside any judging document.
But here's what's really interesting: this is also about elevating our work. If we all have access to winners' data (and they are publicly available), and we can cross-analyze them, why aren't we using it to create better campaigns for clients or write stronger submissions?
Because in the current economy, you can't afford to run dull ads. A study by System1, eatbigfish & Peter Field found that dull TV ads would require $228 billion in extra spending to achieve the same market share growth as the most impactful ads. We also know from Paul Dyson's research that creativity is the #1 factor within our control that influences brand profitability.
The 2025 Cannes Lions Titanium winners will be announced on June 20th. We'll be watching.
The data in this analysis comes from Cannes Lions 2024 winners, aggregated together with 4,000+ more winners from the best award shows spanning 2019-2025. We'll be covering all upcoming Cannes winners in our platform. Visit Adaily and choose the filter on the right to browse Cannes 2024 winners - and 2025 winners soon.
Want to explore the intersection of AI and creativity? This experiment is just the beginning of understanding what machines can teach us about human creative excellence. Follow us for more.