Skip to content
MKTGAGENT.AI MKTGAGENT.AI
Blog

Rand Fishkin proved AI recommendations are inconsistent – here’s why and how to fix it

William Gu
William Gu
Marketing AI Team @ MKTGAGENT.AI

Why Rand Fishkin’s Research Exposes the Inconsistency in AI Brand Recommendations—and What Brands Can Do About It

Artificial intelligence (AI) is increasingly influential in shaping brand visibility through its recommendations. Yet Rand Fishkin’s recent research highlights a critical flaw: AI recommendations for brands are alarmingly inconsistent. This inconsistency challenges traditional ranking metrics and signals a deeper issue with how AI systems determine brand prominence.

The Problem: Inconsistent AI Recommendations

Fishkin’s analysis found that across various AI platforms, identical brand recommendation lists appeared in less than 1% of runs. This unpredictability renders conventional ranking methods ineffective. Why does this happen? The root cause lies in what Fishkin calls the “confidence problem”—how AI gauges trust and reliability in the entities it recommends.

Understanding the “Confidence Problem” and Cascading Confidence

AI systems rely on a pipeline to assess and present information. At each stage, confidence—or trust—is accumulated. Fishkin introduces the concept of “cascading confidence,” which describes how trust builds and flows through these stages. If a brand’s presence or related information is lacking or inconsistent along this chain, the AI’s confidence diminishes, leading to erratic recommendation results.

How Brands Can Improve Visibility

To combat this, Fishkin outlines strategic methods brands can adopt:

  • Optimize the “Entity Home”: This refers to a brand’s primary digital presence, such as its official website or profile pages. Clear, authoritative, and up-to-date information here boosts initial confidence.
  • Corroboration from Independent High-Authority Sources: AI systems place greater trust in entities verified by credible, external sources. Ensuring positive and consistent mentions across respected outlets strengthens a brand’s profile.
  • Presence Across Multiple Knowledge Graphs: Visibility in diverse knowledge graphs—databases that connect and organize information—signals widespread recognition and reliability.

Key Insights

  • Why do AI brand recommendations vary so greatly? It’s due to the “confidence problem” impacting how AI systems trust and verify information.
  • How can brands become more consistently visible to AI? By optimizing their digital presence and securing corroboration from reputable sources.
  • What role do knowledge graphs play? They provide a broad set of verification points that enhance AI confidence.

Conclusion

Fishkin’s research exposes a vital opportunity for brands: as AI becomes central to online recommendation systems, building reliable, consistent signals across the web is no longer optional. By understanding and addressing the “confidence problem,” brands can avoid falling into the low-confidence zone and instead become favored, trustworthy choices in AI-driven spaces. Proactive management of a brand’s digital ecosystem will be key to thriving in the evolving AI landscape.

Source: https://searchengineland.com/ai-recommendations-inconsistent-fix-469250

Tags: