Many so-called AI strategies are actually marketing and PR strategies driven by hype machines. According to Dr. Uri Yerushalmi, chief AI officer and co-founder of Fetcherr, this confusion has reached a tipping point. “The word ‘AI’ has been diluted,” he told me in the interview. "Any application that uses language models now tends to label itself as an AI product. There is a clear difference between companies that use AI tools and companies that really develop proprietary AI technologies."

And in 2025, as budgets tighten and investors demand more substantive content, this distinction may determine who can survive.

The rise and risk of AI whitewashing

The AI ??revival has companies scrambling to integrate ChatGPT, Claude, Gemini and other large language models into their products. But calling this “AI transformation” is like painting racing stripes on an old car and calling it a Ferrari.

Yerushalmi pointed out a systematic mislabeling problem: "Companies and consumers confuse AI with language models. But LLM is only one aspect of the broader AI landscape." In fact, Garter's hype cycle on generative AI predicts that "by 2026, 80% of enterprises will use GenAI API", but using APIs is not the same as building an intelligent system.

When companies exaggerate AI’s statements beyond reality, the consequences are more than just embarrassment. "One of the biggest risks is that companies may start to rely too much on AI, reduce security measures and make decisions without adequate manual supervision," Yerushalmi said. This also undermines the trust of employees, investors and the public.

One example of global headlines is Air Canada’s AI agent, which mistakenly informs customers about bereavement discounts. The judge later ruled that the airline was responsible for its AI's misrepresentation—AirCanada had to cash in on false discounts. The ruling sets a strong precedent: companies are responsible for what their AI says.

Yerushalmi added: "When AI is integrated in a progressive and controllable way, it becomes a powerful decision-making tool. But overcommitment can ruin everything."

How investors identify substance

Not every product that claims to be AI deserves this label. "If a product uses AI in a trivial or superficial way, it's likely just a hype. The real innovation lies in the companies that build proprietary, revolutionary AI that fundamentally reshapes the operational processes of the entire industry," Yerushalmi explained.

At Fetcherr, this innovation is embodied in a real-world system that combines traditionally isolated pricing and revenue management capabilities in the aviation industry. This is a great example of how AI can really refactor workflows, not just automate tasks. It is crucial for investors to identify this distinction.

The essence looks like performance metrics, real-time use cases and product evolution—not just dazzling AI demonstrations or GPT plugins.

Risks faced by brands

Overhyped AI products have multiple consequences for individuals and businesses. As a business, you are at risk not only the risk of technological failure, but also the risk of damage to your brand reputation. According to a 2023 Cisco survey, 60% of consumers are worried about how AI uses their data, and 65% say their trust in their organizations has been lost due to the abuse of AI. In another newer survey by Research Information, nearly every respondent was “concerned that AI would be used to spread false information and could lead to critical errors or accidents.”

These data show that AI and organizations are growing credibility crises, and organizations now need to build and deploy AI tools in a way that maintains trust and enhances credibility, rather than erode it. Yerushalmi reiterated that “the main damage is the loss of trust.” This loss of trust can have catastrophic consequences when AI is deployed at scale in customer interactions and regulated workflows.

How to penetrate noise

So how do companies penetrate noise? For Yerushalmi, the answer lies in the A/B test. "In business, the ultimate goal is to increase profitability and efficiency, and the best way to verify AI performance is to use scientific tools such as A/B testing," he said.

As AI’s influence grows, measurable standards are critical to building effective AI models and deploying them securely and successfully. Activities like MLCommons and Stanford’s HELM benchmarks – both highly praised by many experts in the industry – are designed to provide transparency in model performance, bias and security. For corporate teams, this rigor can distinguish real solutions from speculative ones.

The real AI revolution

While Yerushalmi agrees that we are indeed in the middle of the AI ??revolution, he believes that “soon, models like Fetcherr’s LMM will drive the actual operation of the enterprise.” He said that this transformation is not led by chatbots, but by decision-making engines that efficiently optimize operations, pricing, logistics and strategy.

While what will really bring in the future remains to be seen, the industry generally believes that the real potential of AI is not in eye-catching demonstrations, or even in billions of dollars of investment, although these are great, but at the heart of the embedded intelligence transformation industry.

The above is the detailed content of How To Tell If Your AI Strategy Is Real Or Just Another PR Hype. For more information, please follow other related articles on the PHP Chinese website!