90% of consumer brands face negative AI sentiment: Report

AI assistants such as ChatGPT, Gemini, Claude and Perplexity are increasingly determining which brands customers discover, trust and buy from, Pulp Strategy said in a new study, powered by Neurorank.

Key findings cited in the report include that 68 percent of brands do not appear in AI-generated lists within their own categories; 52 percent face factual errors or misstatements in AI responses; 90 percent of consumer brands show a negative sentiment skew; and 88 percent suffer inconsistent naming and recognition across models.

The GEO Benchmark Index 2025, the agency said, is based on 70 enterprise audits, more than 350 brands, 408,000 prompt simulations and interviews with more than 60 chief marketing officers and other senior marketing executives. The study found a wide gap between what brands communicate and what large language models actually recall.

The report said the structural weaknesses stem from non-machine-readable content, outdated formats such as PDFs and unstructured pages, and weak trust signals in AI knowledge graphs. Those issues, the study argues, can lead to brands being omitted from comparison queries, outdated or fabricated capabilities appearing in answers, and competitors surfacing ahead of category leaders.

Ambika Sharma, founder and chief strategist at Pulp Strategy, said, “AI assistants via AI search are now the gatekeepers of demand. If a brand is missing from AI-generated answers, it is missing from the customer’s world entirely. It's taken us 4+ months and over 100 conversations with CMOs’ leaders across industries to capture nuances and insights coded here. Our biggest learning was that Neurorank is a powerful system that is built for impact in an AI search world with foggy visibility.”

The research also warned of business consequences, saying AI search now shapes procurement shortlists, investor narratives and high-intent consumer discovery. It cited external benchmarks showing a “4,700 percent YoY rise in traffic from generative platforms to retail websites,” and higher buying intent from such traffic.

Pulp Strategy described its methodology as combining 70 anonymised GEO audits, 408,000 real-world prompt simulations across ChatGPT, Gemini, Claude and Perplexity, interviews with CMOs and other professionals, and external validation of select findings.

The report identifies risks for regulated industries, including misinterpreted ESG credentials, compliance exposure and impaired procurement visibility. It recommends actions to improve how brands appear inside AI responses, and offers a 30-90 day roadmap to increase factual accuracy, reduce hallucinations, strengthen AI-trusted content surfaces and improve cross-model recognition.

“AI assistants are rewriting how markets understand brands,” Ambika Sharma said. “This research shows the reality inside the models - and most brands will not like what they find.”