Embedded Files
Click to expand/hide the summary
Business Problem
Balancing authenticity, scalability, and commercial intent in the fast-evolving influencer marketing landscape.
Our Vision
AI-driven models that decode successful influencer engagement patterns, merging authenticity with brand goals for scalable marketing.
How It Works
AI identifies key engagement drivers from social data.
Models generate influencer-style content aligned with brand messaging.
Influencers and brands use AI to scale reach authentically.
Key Benefits
Adaptive, scalable influencer marketing.
Stronger audience-message connection.
AI-powered campaign prototyping.
AI Challenges
Developing "AI creatives" that seamlessly embed commercial intent into authentic influencer messaging.
Business problem
Influencers have become a major force in modern marketing, bridging the gap between brands and audiences with authenticity, relatability, and trust. However, replicating the magic of successful influencers is far from straightforward. Brands often struggle to navigate the delicate balance between authentic engagement and commercial intent.
The stakes are high for brands, as misjudging audience expectations or adopting overly generic approaches can lead to wasted efforts and reputational risks. Meanwhile, influencers themselves face challenges in maintaining their authenticity while scaling their reach and collaborating with brands effectively.
Adding to the complexity, the sheer volume of influencers makes strategic decision-making even more difficult. Achieving broad yet affordable reach requires brands to move beyond celebrity endorsements and tap into the long tail of micro- and nano-influencers. However, this approach comes with its own challenges—these influencers emerge and fade quickly, making influencer marketing a highly dynamic and unpredictable space.
To navigate these obstacles, brands often resort to limited strategies: repeated collaborations with established influencers face diminishing returns, niche targeting lacks scalability, and specialized agencies charge a premium while struggling to keep pace with the rapidly shifting landscape. It is no surprise that influencer marketing ROI can be so inconsistent.
These limitations highlight the pressing need for dynamic, scalable, and cost-effective solutions that can deliver authenticity and impact in influencer marketing.
Approach
Our approach explores the potential of AI to decode and replicate the patterns of influence. By analyzing data from social platforms, audience interactions, and successful campaigns, we aim to build models that understand what drives engagement and trust in specific contexts.
Using these models, we can design AI systems capable of crafting messages and strategies that emulate an influencer’s unique style and resonate with target audiences. For influencers, this technology can act as a supportive tool, helping them understand their audience better and amplify their impact without compromising their voice.
For brands, this means scalable, data-driven methods to create personalized campaigns that feel genuine and authentic. These models could even help brands prototype influencer content before production, supporting processes like influencer search and idea generation.
We aim to develop scalable models able to encompass virtually all influencers in the market and their interactions with audiences, creating a dynamic foundation for brands to execute cost-effective, authentic, and impactful influencer marketing campaigns.
AI challenges
While the current capabilities of Large Language Models (LLMs)—including techniques like fine-tuning—provide a strong foundation for learning patterns of authentic engagement desired by influencers and brands, several challenges remain.
One key challenge is LLM alignment, ensuring the model prioritizes behaviors that drive meaningful engagement over simply replicating frequent patterns. This is critical in avoiding diminishing returns, particularly in marketing contexts. Advanced alignment methods, such as Reinforcement Learning from Human Feedback (RLHF) and AI Feedback (RLAIF), will play a vital role in steering the model toward these goals.
The greater challenge, however, lies in developing "AI creatives"—models capable of embedding a brand’s commercial intent seamlessly into influencer-style messages. This requires advanced problem-solving capabilities at inference time, enabling the model to craft authentic, engaging content that aligns with both the influencer’s voice and the brand’s goals. While most popular LLMs currently fall short of this level of creativity, emerging problem-solving-oriented models hold promise and could potentially pave the way for the development of true AI creatives.
Open roles
There are currently no open roles for this initiative.
Follow us on LinkedIn for updates.
Partnerships
This initiative is conducted in partnership with the Institute of Mathematics and Computer Science (ICMC) of the University of São Paulo (USP).
Partner
Sponsor
Report abuse