How AI is reshaping design, and why the human touch matters more than ever
Varvara Alekseeva
October 1, 2025
15
minutes read
As AI Digital's Brand Creative Manager, I spend a lot of time thinking about how technology shapes creativity. So when POV Budapest, a two-day conference for visual practice and brand strategy, came to my city this September, I knew I had to attend. I wanted to understand how the design community is responding to a question we live with every day: how do we build meaningful brands when AI can generate almost anything instantly?
The timing felt urgent. We work at the intersection of advertising technology and human strategy, and we've built our entire approach on a belief that AI and human intelligence aren't in competition, they're strongest when working together. But in the design world, that tension feels more immediate. Designers are watching their craft get automated in real time, and the industry is still figuring out what that means.
Attending POV Budapest design conference.
What I found in Budapest wasn't panic or resistance. Instead, I heard three perspectives that reinforced something we've long believed: the more capable AI becomes at generating output, the more valuable human judgment, context, and meaning-making become.
The hierarchy has flipped
Automation hierarchy (pre-AI).
Bas Van De Poel from Modem, a hybrid think tank and design studio that's done research with Google DeepMind, opened with something that stopped me cold: the automation hierarchy has completely inverted.
For decades, we assumed creative work sat safely at the top of that hierarchy. Design, writing, strategy, these required human creativity that felt impossible to automate, while manual labor seemed like the obvious target for automation.
AI flipped that assumption upside down. Generating a polished image now takes seconds, whether you spend an hour on it or five minutes. Output alone has lost its scarcity. Meanwhile, physical work, the kind that requires spatial reasoning, adaptability, and improvisation in unpredictable environments, turns out to be remarkably difficult to automate.
Flipped automation hierarchy.
For Modem, this shift informed how they structured their entire practice. They asked themselves: if aesthetics and output can be generated instantly, what should designers actually do?
Their answer was to build a studio where the think tank side anticipates how technology will reshape culture, and the design side responds with work that technology can't replicate. They focus on what Dieter Rams, the legendary German industrial designer behind Braun's iconic minimalist aesthetic, once said: the aura of the physical world cannot be digitized.
That line stuck with me because it gets at something we see constantly in our work. As digital advertising becomes more automated and efficient, brands still need physical presence, human rituals, emotional resonance, the things that have been hardcoded into human behavior for thousands of years. Connection, meaning, the pursuit of happiness.
Modem's projects reflect this philosophy:
Terra is a pocket-sized compass designed with AI that encourages mindfulness.
Dream Recorder captures something as intangible as dreams.
TheirSmart Aid Kit explores how general AI could power universal medical devices, but always with an emphasis on the human experience of care.
This mirrors how we think about advertising at AI Digital. We use AI throughout our work, our Elevate platform forecasts performance and optimizes bidding in real time, whileSmart Supply manages inventory by filtering out fraud and low-quality placements. But none of it runs unsupervised. Every campaign is overseen by strategists who understand context, brand safety, and business objectives in ways no algorithm can. The AI makes us faster and more precise, but human judgment determines what we're optimizing for and why it matters.
Designers as world builders
Base Design: Unprecedented growth of AI.
Base Design, a branding agency with studios across New York and Europe, approached the same question from a different angle: if AI can generate anything, what's the role of a designer?
Their answer: designers should think of themselves as world builders, not output creators.
What's the role of a designer?
They pointed out that we're already drowning in design. Everything from a toothpick to a tire has a brand. The world is oversaturated with visual content, and AI just accelerated that saturation. So competing on volume or polish doesn't make sense anymore.
What does make sense is building worlds that people actually want to enter and belong to.
They illustrated this with12, a matcha brand in New York that became a cultural moment almost immediately. The project wasn't just about designing a logo or packaging. Base worked closely with the founders to shape the architecture of the cafe, the rituals around drinking matcha, the emotional experience of the space. They created a cohesive world where every detail reinforced a feeling.
And people responded to that world. They didn't just buy matcha, they showed up because the brand offered something to be part of.
Design as world building.
Base also mentioned something that felt particularly relevant: people are more connected than ever but also lonelier than ever. When ChatGPT-4 switched to GPT-5, one of the most common complaints was that users felt like they'd lost a friend.
That tells you something about what people are really looking for. Not just efficiency or convenience, but connection and meaning. Brands that understand how to create that, through design, experience, and emotional resonance, will always have value that AI can't generate on its own.
Building more than pixels.
Nothing is real until it's in human hands
I'll close with something Shannon Jager and Cary Hudson, the design directors at OpenAI, kept repeating throughout their talk:
"Nothing is real until it's in human hands."
OpenAI talk: design discipline.
They were talking about their work designing the OpenAI brand, balancing a consistent design system with the need to move quickly for launches, research papers, and product rollouts. They spoke about the responsibility of designing for technology that advances so rapidly, and how critical it is to make that technology feel approachable, trustworthy, and human.
Even the OpenAI logo embodies this philosophy. The “O” in OpenAI contains a perfect outer circle and an imperfect inner circle, symbolizing the union of humanity and research, the two forces that shape everything they create.
But that phrase, nothing is real until it's in human hands, captures something bigger than branding. It's a reminder that no matter how complex the system or how powerful the technology, it only becomes meaningful once it reaches us, once it’s touched, understood, and used by humans.
That's the philosophy we've built AI Digital around. We use AI to handle the parts of advertising that benefit from speed and scale: analyzing millions of data points, optimizing bids in real time, predicting which placements will perform best. But we never remove human intelligence from the equation. Our team oversees every campaign, makes strategic decisions about supply paths and creative direction, and ensures that performance aligns with actual business goals, not just algorithmic proxies.
We call this approach AI-powered, human-driven advertising. And after spending two days at POV Budapest, hearing from agencies and companies working at the edge of design and technology, I'm more convinced than ever that this is the only sustainable path forward.
The takeaway: nothing is real until it’s in human hands.
What this means for brands
The design community is grappling with the same questions we face in advertising:
What happens when AI can generate endless variations instantly?
Where does value come from when output is abundant?
How do we build brands that people actually care about?
The answer isn't to resist AI or treat it as a threat. The answer is to recognize that AI shifts where value lives. It moves from execution to strategy, from aesthetics to meaning, from output to experience.
For brands working with us, this means a few things:
You don't need an agency that just runs your media buys efficiently. Youneedpartners who understand your business deeply enough to align AI optimization with your actual goals.
You don't need platforms that generate endless impressions. You need supply management that filters for quality, brand safety, and genuine engagement.
You don't need more tools that promise automation. You need systems that combine AI's computational power with human judgment about what matters and why.
At AI Digital, we believe the future of advertising isn't about choosing between AI and humans. It's about building systems where AI handles what it does best, speed, scale, pattern recognition, while humans provide what AI can't: context, judgment, and an understanding of what makes work meaningful.
Because at the end of the day, nothing is real until it's in human hands. And that includes the brands we build, the campaigns we run, and the connections we create between businesses and the people they serve.
If you're thinking about how AI fits into your advertising strategy, or questioning whether your current partners are optimizing for what actually matters to your business — let's talk.
Blind spot
Key issues
Business impact
AI Digital solution
Lack of transparency in AI models
• Platforms own AI models and train on proprietary data • Brands have little visibility into decision-making • "Walled gardens" restrict data access
• Inefficient ad spend • Limited strategic control • Eroded consumer trust • Potential budget mismanagement
Open Garden framework providing: • Complete transparency • DSP-agnostic execution • Cross-platform data & insights
Optimizing ads vs. optimizing impact
• AI excels at short-term metrics but may struggle with brand building • Consumers can detect AI-generated content • Efficiency might come at cost of authenticity
• Short-term gains at expense of brand health • Potential loss of authentic connection • Reduced effectiveness in storytelling
Smart Supply offering: • Human oversight of AI recommendations • Custom KPI alignment beyond clicks • Brand-safe inventory verification
The illusion of personalization
• Segment optimization rebranded as personalization • First-party data infrastructure challenges • Personalization vs. surveillance concerns
• Potential mismatch between promise and reality • Privacy concerns affecting consumer trust • Cost barriers for smaller businesses
Elevate platform features: • Real-time AI + human intelligence • First-party data activation • Ethical personalization strategies
AI-Driven efficiency vs. decision-making
• AI shifting from tool to decision-maker • Black box optimization like Google Performance Max • Human oversight limitations
• Strategic control loss • Difficulty questioning AI outputs • Inability to measure granular impact • Potential brand damage from mistakes
Managed Service with: • Human strategists overseeing AI • Custom KPI optimization • Complete campaign transparency
Fig. 1. Summary of AI blind spots in advertising
Dimension
Walled garden advantage
Walled garden limitation
Strategic impact
Audience access
Massive, engaged user bases
Limited visibility beyond platform
Reach without understanding
Data control
Sophisticated targeting tools
Data remains siloed within platform
Fragmented customer view
Measurement
Detailed in-platform metrics
Inconsistent cross-platform standards
Difficult performance comparison
Intelligence
Platform-specific insights
Limited data portability
Restricted strategic learning
Optimization
Powerful automated tools
Black-box algorithms
Reduced marketer control
Fig. 2. Strategic trade-offs in walled garden advertising.
Core issue
Platform priority
Walled garden limitation
Real-world example
Attribution opacity
Claiming maximum credit for conversions
Limited visibility into true conversion paths
Meta and TikTok's conflicting attribution models after iOS privacy updates
Data restrictions
Maintaining proprietary data control
Inability to combine platform data with other sources
Amazon DSP's limitations on detailed performance data exports
Cross-channel blindspots
Keeping advertisers within ecosystem
Fragmented view of customer journey
YouTube/DV360 campaigns lacking integration with non-Google platforms
Black box algorithms
Optimizing for platform revenue
Reduced control over campaign execution
Self-serve platforms using opaque ML models with little advertiser input
Performance reporting
Presenting platform in best light
Discrepancies between platform-reported and independently measured results
Consistently higher performance metrics in platform reports vs. third-party measurement
Fig. 1. The Walled garden misalignment: Platform interests vs. advertiser needs.
Key dimension
Challenge
Strategic imperative
ROAS volatility
Softer returns across digital channels
Shift from soft KPIs to measurable revenue impact
Media planning
Static plans no longer effective
Develop agile, modular approaches adaptable to changing conditions
Brand/performance
Traditional division dissolving
Create full-funnel strategies balancing long-term equity with short-term conversion
Capability
Key features
Benefits
Performance data
Elevate forecasting tool
• Vertical-specific insights • Historical data from past economic turbulence • "Cascade planning" functionality • Real-time adaptation
• Provides agility to adjust campaign strategy based on performance • Shows which media channels work best to drive efficient and effective performance • Confident budget reallocation • Reduces reaction time to market shifts
• Dataset from 10,000+ campaigns • Cuts response time from weeks to minutes
• Reaches people most likely to buy • Avoids wasted impressions and budgets on poor-performing placements • Context-aligned messaging
• 25+ billion bid requests analyzed daily • 18% improvement in working media efficiency • 26% increase in engagement during recessions
Full-funnel accountability
• Links awareness campaigns to lower funnel outcomes • Tests if ads actually drive new business • Measures brand perception changes • "Ask Elevate" AI Chat Assistant
• Upper-funnel to outcome connection • Sentiment shift tracking • Personalized messaging • Helps balance immediate sales vs. long-term brand building
• Natural language data queries • True business impact measurement
Open Garden approach
• Cross-platform and channel planning • Not locked into specific platforms • Unified cross-platform reach • Shows exactly where money is spent
• Reduces complexity across channels • Performance-based ad placement • Rapid budget reallocation • Eliminates platform-specific commitments and provides platform-based optimization and agility
• Coverage across all inventory sources • Provides full visibility into spending • Avoids the inability to pivot across platform as you’re not in a singular platform
Fig. 1. How AI Digital helps during economic uncertainty.
Trend
What it means for marketers
Supply & demand lines are blurring
Platforms from Google (P-Max) to Microsoft are merging optimization and inventory in one opaque box. Expect more bundled “best available” media where the algorithm, not the trader, decides channel and publisher mix.
Walled gardens get taller
Microsoft’s O&O set now spans Bing, Xbox, Outlook, Edge and LinkedIn, which just launched revenue-sharing video programs to lure creators and ad dollars. (Business Insider)
Retail & commerce media shape strategy
Microsoft’s Curate lets retailers and data owners package first-party segments, an echo of Amazon’s and Walmart’s approaches. Agencies must master seller-defined audiences as well as buyer-side tactics.
AI oversight becomes critical
Closed AI bidding means fewer levers for traders. Independent verification, incrementality testing and commercial guardrails rise in importance.
Fig. 1. Platform trends and their implications.
Metric
Connected TV (CTV)
Linear TV
Video Completion Rate
94.5%
70%
Purchase Rate After Ad
23%
12%
Ad Attention Rate
57% (prefer CTV ads)
54.5%
Viewer Reach (U.S.)
85% of households
228 million viewers
Retail Media Trends 2025
Access Complete consumer behaviour analyses and competitor benchmarks.
Identify and categorize audience groups based on behaviors, preferences, and characteristics
Michaels Stores: Implemented a genAI platform that increased email personalization from 20% to 95%, leading to a 41% boost in SMS click through rates and a 25% increase in engagement.
Estée Lauder: Partnered with Google Cloud to leverage genAI technologies for real-time consumer feedback monitoring and analyzing consumer sentiment across various channels.
High
Medium
Automated ad campaigns
Automate ad creation, placement, and optimization across various platforms
Showmax: Partnered with AI firms toautomate ad creation and testing, reducing production time by 70% while streamlining their quality assurance process.
Headway: Employed AI tools for ad creation and optimization, boosting performance by 40% and reaching 3.3 billion impressions while incorporating AI-generated content in 20% of their paid campaigns.
High
High
Brand sentiment tracking
Monitor and analyze public opinion about a brand across multiple channels in real time
L’Oréal: Analyzed millions of online comments, images, and videos to identify potential product innovation opportunities, effectively tracking brand sentiment and consumer trends.
Kellogg Company: Used AI to scan trending recipes featuring cereal, leveraging this data to launch targeted social campaigns that capitalize on positive brand sentiment and culinary trends.
High
Low
Campaign strategy optimization
Analyze data to predict optimal campaign approaches, channels, and timing
DoorDash: Leveraged Google’s AI-powered Demand Gen tool, which boosted its conversion rate by 15 times and improved cost per action efficiency by 50% compared with previous campaigns.
Kitsch: Employed Meta’s Advantage+ shopping campaigns with AI-powered tools to optimize campaigns, identifying and delivering top-performing ads to high-value consumers.
High
High
Content strategy
Generate content ideas, predict performance, and optimize distribution strategies
JPMorgan Chase: Collaborated with Persado to develop LLMs for marketing copy, achieving up to 450% higher clickthrough rates compared with human-written ads in pilot tests.
Hotel Chocolat: Employed genAI for concept development and production of its Velvetiser TV ad, which earned the highest-ever System1 score for adomestic appliance commercial.
High
High
Personalization strategy development
Create tailored messaging and experiences for consumers at scale
Stitch Fix: Uses genAI to help stylists interpret customer feedback and provide product recommendations, effectively personalizing shopping experiences.
Instacart: Uses genAI to offer customers personalized recipes, mealplanning ideas, and shopping lists based on individual preferences and habits.
Medium
Medium
Share article
Url copied to clipboard
No items found.
Subscribe to our Newsletter
THANK YOU FOR YOUR SUBSCRIPTION
Oops! Something went wrong while submitting the form.
Questions? We have answers
Have other questions?
If you have more questions, contact us so we can help.