Summary of Key Points
Artificial Intelligence (AI) is profoundly changing the way consumers make purchasing decisions (nearly 80% of users consult AI before placing orders). The focus of brand competition has shifted from being “seen by consumers” to being “selected by AI.” Traditional brand communication systems are becoming ineffective; brand images are fragmented and difficult to control within the AI ecosystem, and brands also face new risks such as cognitive contamination. A report from Fudan University proposes the “AI Brand Health Index” (a three-tier structure: C-B-A), which guides companies to start with their underlying digital assets and transition to long-term “cognitive asset management.” However, small and medium-sized enterprises (SMEs) still face challenges in implementing these strategies.
1. Nearly 80% of Consumers Consult AI Before Making Purchases: The Landscape of Buying Decisions Has Changed
In the past, when buying something, you might search for keywords, compare product lists, or read reviews. But now, approximately 80% of people use AI tools like DouBao or QianWen to ask questions like, “Is this brand of phone worth buying?” or “Which skincare product is suitable for sensitive skin?” AI has become the new “super shopping assistant,” significantly shortening the recommendation process. QuestMobile data shows that DouBao has 382 million active users, which means one in every four Chinese people uses it. This change indicates that whether a brand is recommended by AI directly determines whether it will be noticed by consumers.
2. The Rules of Brand Competition Have Changed: From “Traffic Competition” to “Cognitive Asset Management”
What drove brand competition in the past? Spending money on search rankings, information flow advertising, and hiring influencers to promote products—all aimed at gaining “traffic.” Now, with AI as the intermediary, these traditional methods are no longer effective. AI recommendations are based on the information it has “learned” about a brand, not on the amount of advertising a company has spent. Moreover, different AI models (e.g., DouBao and QianWen), at different times, and in different contexts, may provide completely different descriptions of the same brand. For example, if you sell coffee, AI might say your product has a good cost-performance ratio one moment and then describe it as having average quality the next. Companies have no control over the data used by AI or how it portrays them. Therefore, brand competition has evolved into “cognitive asset management”—companies must manage all the information that AI can learn about their brand to make it appear reliable and worthy of recommendation.
3. New Risks for Brands in the AI Era: Cognitive Contamination and Ecosystem Fragmentation
In addition to losing control of their brand image, brands face two major issues:
1. Cognitive Contamination: Some people sell software that can manipulate AI recommendations (for a few hundred yuan) or post negative articles (for a few dozen yuan each, hundreds per day) to intentionally provide false information about a brand. Once this negative information enters the AI system, it is difficult to remove. Unlike web pages, AI updates its knowledge slowly, and the negative information can continue to affect recommendations.
2. Ecosystem Fragmentation: Different AI platforms use different content sources. For example, DouBao may access content from platforms like Douyin and Toutiao, while only YuanBao can access articles from WeChat official accounts. Companies need to create high-quality content on multiple platforms to ensure their brands are recommended by various AI systems, which is a challenging task.
4. How to Make AI “Like” Your Brand? The Report Offers Three Solutions
The Fudan University team proposes the “AI Brand Health Index,” which follows a three-tier structure (C-B-A):
- C Layer (Asset Health): Your public digital assets (website, press releases, product descriptions) should be comprehensive, consistent, and verifiable. For instance, if your website claims your products are organic, your press releases should not contradict this information; otherwise, AI will be confused.
- B Layer (Mechanism Health): Your content should align with AI’s “reading habits.” AI prefers content that is well-structured, logical, and data-driven (e.g., stating “99% of our products meet quality standards” is more useful than simply saying “Our products are of high quality.”
- A Layer (Performance Health): Evaluate how well your brand is actually recommended by AI. For example, when someone searches for “the best milk brand” on DouBao, should your brand appear in the results, and in which position?
Research demonstrates that these three layers are interrelated; poor underlying assets will inevitably lead to poor performance. Companies must start with their basic digital assets and optimize them step by step, rather than focusing solely on crisis management.
5. Implementation Challenges: SMEs Lack the Resources, and Tool Effectiveness Needs to Be Proven
While this approach sounds promising, its implementation is not easy:
- Limited Resources for SMEs: Small companies lack the funds and manpower to manage digital assets across multiple platforms.
- Unreliable Tools: The evaluation tools for AI brand health are still in their early stages, and their accuracy needs to be verified through more practical applications.
Experts emphasize that brand management in the AI era is a long-term effort that cannot be accomplished in one go. Companies should start researching AI recommendation algorithms and optimizing their presence on various AI platforms to avoid being left behind.
In summary, consumers are now relying on AI for their purchasing decisions. Brands must not simply decide whether to adopt AI-based strategies but must determine where to begin—by organizing their digital assets and gradually building trust with AI.