How brands are navigating the rules of the road for AI product recommendations

The rise of AI language models as product recommenders has created a new frontier for brand perception and marketing. Companies are discovering that artificial intelligence systems can significantly influence consumer choices through their product recommendations and interpretations of brand messaging.

The big picture: AI models are increasingly being used to recommend products to consumers, with 28% of people already using AI for product recommendations in areas like cosmetics.

• Companies like JellyFish are developing tools to analyze how different AI models perceive and describe brands
• Brand perception by AI can vary significantly between different models, such as Meta’s Llama and OpenAI’s ChatGPT
• Early evidence suggests that companies can influence how AI models view their products through strategic content creation

Real-world implications: Companies are already experiencing the impact of AI perceptions on their brand messaging and working to adjust their strategies accordingly.

• A meal prep company discovered AI was describing their service as “complicated” simply because their marketing showed chopped chives in food photos
• Ballantine’s whisky found that AI models were incorrectly positioning their mass-market product as premium, prompting them to adjust their marketing assets
• The effectiveness of these adjustments is still being evaluated, with companies reporting early positive results

Technical considerations: The relationship between brands and AI models is becoming increasingly sophisticated, with new tools and methodologies emerging.

• Share of Model software analyzes multiple AI models’ perspectives on brands through automated questioning
• Reasoning models are providing insight into AI decision-making processes by showing their “chain of thought”
• Research has shown that specific word choices in prompts can dramatically influence AI recommendations

Emerging challenges: The intersection of AI and brand marketing is creating new concerns about manipulation and bias.

• Research indicates AI models show inherent bias toward global brands over local ones
• Models tend to associate luxury brands with high-income countries and non-luxury brands with low-income countries
• Companies may attempt to influence AI recommendations through strategic prompt engineering or forum manipulation
• AI companies are likely to develop countermeasures against attempts to manipulate their models’ recommendations

Strategic adaptations: Brands are developing new approaches to maintain positive AI perception while ensuring marketing consistency.

• Companies are using AI as a pre-launch focus group for marketing campaigns
• Brand consistency across different platforms and messages is becoming increasingly important for AI interpretation
• Marketing teams are learning to balance authentic brand messaging with AI-friendly content

Looking ahead: The emergence of AI as a crucial audience for brand messaging represents a fundamental shift in marketing strategy that will require careful navigation of ethical boundaries and technological capabilities. While the potential for manipulation exists, successful brands will likely be those that maintain authenticity while optimizing their presence for AI understanding, rather than those attempting to game the system through deceptive practices.