5 Shocking Truths Investors Can’t Afford to Ignore About AI Productivity

AI’s not the problem. The real problem is how we talk about and measure AI productivity.

Author

Glenda Carnate
January 22, 2026

Read time: 2.5 minutes

If AI-enabled productivity had existed before now, it would not be necessary to explain it; everyone would see its impact on company profits.

During an earnings call, a company's listing of AI initiatives gives executives an air of confidence. As executives discuss ways to increase efficiency through automation and future leverage, investors listen attentively and begin buying stock, raising their expectations for the future. Although some items in the presentation imply productivity, the topic is generally not stated.

However, several months after the first quarter, many company's margins look familiar, the number of employees in the workplace has not increased, the output per employee has not increased, and while technology may be improving at a rapid pace, the way in which the financials are recorded for each employee will not change until investors have to examine a question: Was the supposed productivity potential really a near-term occurrence?

5 Harsh Realities of AI Productivity for Investors:

  1. Productivity should be tangible in the profit margins, but they're nonexistent.
    Corrective Action: Request pre/post AI Unit Economics analysis (money) as proof of productivity. If there's no margin difference, there's no productivity.

  2. Saving time doesn't equate to productivity, but it gives you more time to do.
    Corrective Action: Request verification of FTE decreases and increased output per FTE, not just on word of mouth.

  3. Friction occurs when AI is introduced, but it will eventually remove labour after the entire implementation period.
    The multiple layers of review before you implement AI will rob your return on investment.
    Corrective Action: Only support the development of AI that will eliminate review layers, not just support.

  4. If you don’t rethink how your company works, AI is just an expensive gadget collecting dust.
    AI can’t change how people act... only better incentives can.
    Corrective Action: Request clarification on how a company's roles, quotas, and incentive systems are modified due to the implementation of AI.

  5. There has been a promise of productivity increase before the reality has caught up with the hype.
    Corrective Action: Invest in AI as a process transformation, not a quick means to gain leverage.

💡Key Takeaway: 

The misconception that AI is a failure in productivity stems from the expectation that productivity improvements would come before organisations are forced to evolve their current understanding of how work is performed. Technology evolves quickly, while the systems of incentives, organisational structures and labour economics evolve more slowly.

👉 LIKE this post if you’ve questioned AI productivity claims on earnings calls.

👉 SUBSCRIBE now to receive clear-eyed analysis on the technology driving our economy and incentivizing economic activity.

👉 Follow Glenda Carnate for the most relevant signals amidst all of the noise regarding AI, the financial markets, and the future of work.

👉 COMMENT with the strongest or weakest AI productivity claim you’ve seen.

👉 SHARE this with someone who thinks AI margins should already be obvious.

Reply

or to participate.