News

When an AI answer puts revenue at risk

In March 2025, a customer questioned Wolf River Electric about a troubling claim in Google’s AI Overview. The answer said the solar installer faced a lawsuit from the Minnesota Attorney General over deceptive sales practices. According to Wolf River’s complaint against Google, its CEO explained that the claim was false.

The customer nevertheless canceled a $150,000 contract.[1]

The answer included references, giving the customer sources that appeared to support its conclusions. Yet the complaint alleged that none supported the accusation against Wolf River. One cited source was an attorney general announcement about legal action against other companies. The announcement names those defendants. Wolf River Electric is not among them.[1]

The reported cancellation remains the company’s allegation, but the case illustrates an important distinction.

Being known to AI is not the same as being represented accurately.

An organization can appear in an answer, accompanied by seemingly credible references, and still be given an unsupported reputation that discourages a purchase.[1]

Nor can this exposure be assessed through a single question to a single assistant. The AI ecosystem spans dozens of providers and hundreds of models. Reasoning settings and access to live web information add further differences in how answers are produced. Understanding the exposure requires research across these configurations, not an isolated spot check.[2]

That is the work Modulus does. Our senior researchers investigate how organizations are represented across AI providers, models and settings, preserve the evidence, carry out corrective action and test whether it works. We track changes over time and help organizations address the inaccurate answers that can undermine trust and turn customers away.[3]

Research shows AI models favor authoritative sources.[4] MOD 1401:2026 provides independently reviewed, evidence-backed organizational records designed to help AI distinguish verified facts from unsupported claims.

Sources

  1. [1]PPC Land • Wolf River Electric v. Google, complaint (opens in a new tab)Complaint against Google LLC, published by PPC Land • November 2025
  2. [2]OpenRouter • Models and providers (opens in a new tab)Directory of AI providers and models
  3. [3]Modulus • Home (opens in a new tab)modulus.ai
  4. [4]ACL 2026 • Whose Facts Win? LLM Source Preferences under Knowledge Conflicts (opens in a new tab)Schuster, Gautam and Markert, Heidelberg University and Heidelberg Institute for Theoretical Studies • Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics, pp. 29430–29459 • July 2026

Complimentary 30-Day Evaluation

Give AI an authoritative account of your organization, backed by evidence and maintained by senior AI consultants. Modulus will show you how AI represents your organization and how verification improves its answers. Investigate, review the evidence, perform A/B testing, ask questions, and evaluate the results.

Experience Modulus for 30 days
Request a confidential consultation