The one sure thing about AI is that it’s in a constant state of flux. This is exciting, but it can also make it hard to predict what the coming weeks might hold, let alone the coming year.
That said, we’re fond of futurecasting. To close out a productive and dynamic 2024, we asked 10 experts from within the Stagwell network to share their predictions for what 2025 has in store for AI and marketing.
Read on for what’s inspiring their imaginations—from website UX that morphs in real-time, to ad campaigns where GenAI “manages 90% of execution, with human marketers acting as orchestrators.”
“If 2023 was the year of efficiency, and 2024 was the year of competition, 2025 will be the year of applications—the year people begin to recognize what AI can do and start to build applications around it.
Already tech companies are buying and installing chips at data centers, drawing in massive computing power, and generating new AI applications currently in their infancy.
Agentic AI will change creative workflows by introducing autonomous decision-making agents, capable of targeting audiences with higher efficiency and serving as the chief brand ambassadors for companies.
These agents will personify the brand—each brand will need a personality just as it needs a logo or website. These personalities, whether they are snarky, angelic, or honest agents, will be the primary consumer interface for brands.”
“This past year saw a serious proliferation of content being made using generative AI. Next year, I anticipate a renewed focus on the quality of AI-generated assets. That’s why I’m most excited to keep an eye on AI-driven content creation, and evaluation, at scale.
In 2025, AI will transform content creation and optimization, enabling marketers to scale dynamic, highly relevant creative in real time. AI will autonomously generate multimedia content—like blog posts, videos, and social updates—while also continuously analyzing engagement data to adjust copy, imagery, format, and timing for maximum impact.
In particular, SmartAssets from Stagwell Marketing Cloud will continue to address this challenge by automating content creation, leveraging real-time performance coupled with proprietary behavioral data to drive the brand's unique content strategy.”
“AI can personalize experiences dynamically, and I foresee this facet evolving in 2024. Consider a typical e-commerce site. AI-powered personalization can adjust backgrounds and content in real time to optimize user conversion.
But the process doesn't stop there—think of it as personalization on steroids. AI is allowing us to imagine a scenario in which every aspect of the digital experience adapts to the specific user, from the content to the UX and UI, all based on what they do on the platform, how they engage, how they navigate.
Imagine a website that literally morphs around the user’s needs. If someone visits after engaging with certain content or using specific search terms, the entire experience—content, architecture, and interface—changes to align with their preferences and behavior.
This all requires rethinking traditional workflows. We have to flip the paradigm completely—we have to have the engineers building models, and expressions of those models, so that they can show the designers what’s possible. And then the designers have to push the capabilities of the technology.”
“The main area of growth I’ve been closely tracking is in the realm of image and video generation, with a specific focus on generative adversarial networks (GANs), variational autoencoders (VAEs), vision transformers, and the widely adopted stable diffusion models.
While these advances are transforming multiple industries, the capability for perfect image or video editing through AI remains a work in progress.
By leveraging performance metrics from advertisements alongside classical machine learning, we can envision a future where even “final” creatives can be optimized further, refining their impact and resonance with audiences.
This optimization process would still involve a human-in-the-loop approach to ensure the quality and appropriateness of changes.
However, the ability to edit creatives down to a granular level is likely to see significant progress in the coming year. When combined with agent-based technologies, this approach opens exciting possibilities.
At SmartAssets, for instance, we could ‘agentify’ the platform, enabling specialized agents to handle tasks like creative analytics and image optimization autonomously. This would allow for a more scalable and efficient workflow, where agents collaborate seamlessly with human oversight to deliver truly optimized content.”
“Stagwell in 2025 will move from building AI efficiency tools and capabilities built on existing hyperscale AI LLM models, to pushing the boundaries and building custom solutions which differentiate in the marketing and advertising realm.
Stagwell feels that AI is coming to a watershed moment where you can start getting creative with how you want to use Gen AI and a combination of multimodal channels and technologies to create compelling consumer experiences.
The human element will continue to be enhanced with effective AI, and efficiency will continue to drive optimization across the marketing landscape.”
“I’m intrigued by the fact that anything could be written using AI, and that people are starting to reverse engineer communications as a result. This is an area where I think we may see interesting developments in 2025. Someone may start with AI writing full bodies of text and then add mistakes or human touches to the copy to ensure it’s not too ‘inhuman.’
There are already platforms that exist to flag ‘AI-generated’ writing—and that then offer to rewrite it in a more human voice... using AI, I’d guess.
I love the idea of using AI to write ‘perfect’ copy, and then using another layer of AI to make that copy less perfect. The supposedly human touch—aka, the imperfections—is what we’re willing to pay money for.”
“In 2025, brands need to be thinking about the rise of the AI-driven creative agency.
AI is enabling agencies to produce high-quality video content faster, more efficiently, and with unprecedented scalability. Agencies specializing in AI (like The Dor Brothers) are leading this charge. They’ve harnessed AI technology to produce viral videos that have achieved hundreds of millions of views.
What do AI-driven agencies offer? There’s efficiency at scale—streamlining the production process, letting agencies create multiple variations of videos in record time. Then there’s enhanced creativity, creating visuals and effects that were once impossible or out-of-budget. And then there’s trend responsiveness, with AI opening up real-time adaptation to social media trends.”
“In 2025, the future of marketing and AI is set to revolve around personalized, realtime campaign execution driven by advances in generative AI and adaptive analytics. Tools will evolve to bridge content creation, strategy, and performance tracking seamlessly.
For example, marketers will lean heavily on AI to generate not just assets but entire campaign frameworks that adapt dynamically to audience responses.
Generated video will become a cornerstone of this transformation, with AI producing context-aware video content at scale, tailored to specific audiences, platforms, and trends.
Marketing teams will pivot from creators to curators, overseeing AI systems that continuously refine tone, visuals, and messaging to align with shifting consumer behavior. By late 2025, we might see campaigns where AI manages 90% of execution, with human marketers acting as orchestrators.”
“Gen-AI ads are going to become ubiquitous, and most consumers won't be able to tell the difference (or care much). The cost saving and scaled production possibilities will be too attractive for brands to ignore. Even those more risk-averse brands will start to dabble in AI-ads, and will also start investing in middleware or mechanisms for maintaining brand safety, that make the use of gen-AI content less risky.
Agencies will be asked if they used AI to create content, even if they didn't—with some clients wanting to know why, and some why not.
Either way every time there's a quality concern, many people will jump to the conclusion that AI was to blame.
On the other hand, using AI for more than generating content will help to offset potential errors, as it allows us to mine and combine bigger datasets and better understand consumer behavior.”
“I get a lot of questions about how synthetic influencers may play a role in influencer marketing. A year ago, I was mixed. The pros—cost-effectiveness and minimal PR risks associated with a fake human with no personal history to worry about—are compelling. The cons include audience skepticism, lack of authenticity, and, of course, the singularity consideration.
With the recent release of the Open AI o1 reasoning model, I’m reconsidering my stance on synthetic influencers. I thought AI influencers were little more than chatbots with faces, but I’m changing my tune.
I believe synthetics can be valuable in customer service—personalizing chatbots, supporting sales presentations, producing videos, and even fronting product, as long as it’s clear they are synthetic talent. This use case will continue to grow in 2025. However, I’m also now more bullish on synthetics playing a larger role in brand strategy.
Compared to the 400 million influencers worldwide, synthetic creations are still a minority. For now, brand relationships will likely be limited to products where avatars are most effective — such as fashion, rather than food or travel — as well as for pitches, A/B testing, or for enabling passive income for influencers who approve their likeness for use.
However, this dynamic will shift as creating synthetic influencers becomes easier and consumers grow more open to following them. At the same time, using real-life talent and creators will become increasingly valuable, distinguishing high-end, authentic brands from more mass-market, commodified brands.”
