If you played a drinking game where you took a shot every time a CEO said "AI" during an earnings call this season, you wouldn't make it past the opening remarks.

"Artificial Intelligence" has become the new "Cloud," the new "Crypto," and the new "Dot Com." It is the mandatory buzzword of the decade. Every company, from massive tech hyperscalers to sleepy regional banks, is suddenly claiming to be an "AI-first organization."

For investors, this creates a dangerous noise-to-signal ratio.

If everyone claims to be an AI winner, nobody is. To find the true beneficiaries—and avoid the pretenders—you need to look beyond the keyword count. You need to analyze the Context of Adoption.

Here is how to use executive commentary to separate the builders from the tourists.

‍

The "Hype vs. CapEx" Ratio

The easiest way to spot a "Fake AI" company is to look for the mismatch between their mouth and their wallet.

We track a metric called the Hype/CapEx Ratio.

• The Hype: How many times did management mention "AI," "GenAI," or "Machine Learning" in the transcript?
• The CapEx: Did they explicitly mention increasing Capital Expenditures (CapEx) or R&D budgets to pay for it?

The Red Flag: A company that mentions AI 50 times but guides for flat R&D spending.

If they aren't buying GPUs, hiring data scientists, or paying for cloud inference, they aren't doing AI. They are just doing marketing.

The Green Flag: A company that warns of "short-term margin compression due to accelerated infrastructure investment." This sounds negative, but it is the truest signal of actual adoption. It means they are putting real money on the line.

‍

Analyzing the "Verb" Selection

To gauge the maturity of a company's AI strategy, look at the verbs they attach to the noun "AI."

Our semantic analysis reveals three distinct stages of adoption across the S&P 500:

1. The "Exploring" Phase (Low Conviction):
   
   • Keywords: "Exploring," "Assessing," "Piloting," "Studying."
   • Verdict: They are reacting to FOMO. No revenue impact for 2+ years.
2. The "Integrating" Phase (Medium Conviction):
   
   • Keywords: "Embedding," "Enhancing," "Optimizing," "Co-pilot."
   • Verdict: They are adding AI wrappers to existing products. Good for efficiency, but not a new business line.
3. The "Transforming" Phase (High Conviction):
   
   • Keywords: "Scaling," "Deploying," "Monetizing," "Inference."
   • Verdict: This is the real deal. They have moved from R&D to production.

By filtering your search results by these verb clusters, you can instantly rank a sector from "Leaders" to "Laggards."

‍

The "Efficiency" Bluff

A common narrative right now is: "AI will replace 30% of our back-office costs."

CEOs love this story because it promises higher margins without revenue growth. But you need to verify it.

The Test: If a CEO claims AI is driving efficiency, search the transcript for "Headcount" and "Hiring."

• If they claim AI efficiency but are still growing non-engineering headcount by 10%, they are lying to themselves (and you).
• True AI adoption in operations should show up as "operating leverage"—revenue growing faster than opex. If the opex line isn't bending, the AI isn't working yet.

‍

The "Derivative" Trade

Finally, the smartest way to play AI adoption isn't just buying the tech stocks; it's finding the Industrial Derivatives.

Use Concept Search to look for "AI" mentions in unexpected sectors:

• Utilities: Who is talking about "Data Center Power Demand"?
• Real Estate (REITs): Who is talking about "Cooling requirements" or "Server floor weight"?
• Industrials: Who is talking about "Copper demand" for grid upgrades?

These companies aren't building LLMs, but they are selling the pickaxes and shovels (or electricity and air conditioning) to the people who are. Their commentary is often far more honest because they treat AI as a customer demand, not a stock promotion.

‍

Don't Buy the Buzzword

The market rewards execution, not aspirations.

In 2024, "We have an AI strategy" is table stakes. It means nothing.

"We are spending $500M on inference clusters to drive $1B in new ARR" is a thesis.

Use the transcript to ignore the former and find the latter.

‍

Who is actually building?

Nextmark’s Concept Search allows you to filter AI mentions by "CapEx Intent" and "Revenue Impact" to find the real beneficiaries.

‍