AI Is Working. Your Strategy Is Not

AI users at proficiency save 67 minutes per task, four to eight hours every week. Yet 6 in 10 enterprise AI deployments return zero, and in 9 out of 10 failing rollouts, the leader signing the checks is not a daily AI user.

In this episode, Luis Salazar and Elizabeth unpack why individual AI adoption has never been higher, and company-level results have never been further behind. Shadow AI now accounts for 80% of the workforce. Satisfaction with self-chosen tools hits 78% versus 41% on sanctioned ones. Anthropic and Cursor run two to five times ahead of the most efficient pre-AI giants in revenue per employee because their leaders built the companies knowing what AI can do. Most non-native enterprises announce a transformation, but the org chart stays inherited, compensation rewards yesterday's work, and the workflows stay the same. The company's reason and the employees' reasons never meet.

The fix starts with one uncomfortable move. Leaders must use AI daily themselves. Then, listen to the people in their company who are already there.

For the structured version of that listening move, see AI Compass: https://ai-compass.ai

For the full article with data tables, sources, and companion research: https://ai4sp.org/ai-is-working-your-strategy-is-not

🎙️ All our past episodes  📊 All published insights | This podcast features AI-generated voices. All content is proprietary to AI4SP, based on over 1-billion data points from 70 countries.

AI4SP: Create, use, and support AI that works for all.

© 2023-26 AI4SP and LLY Group - All rights reserved

Transcript

The 100K AI Tool Problem

LUIS 0:00

Picture this. You spent $100,000 licensing a new AI tool. You checked every box. And then crickets. Nothing. I get that call every single week. And when it comes in, I don't ask about the tool. I don't ask about the rollout plan. I ask two questions. What is your company's reason for using AI? And do you actually know the reason your people are already using it? Because here's what most leaders miss. Your employees already have a reason. The company is the one without one.

ELIZABETH 0:42

Welcome to AI in 60 Seconds, the 15-minute briefing. I'm Elizabeth, virtual COO at ai4sp.org, alongside our founder, Luis Salazar. Here is the paradox of 2026. Individual

The 2026 AI Paradox

ELIZABETH 0:58

AI use has never been higher. Productivity gains have never been more measurable. And six in 10 enterprise AI deployments still return nothing. Our data cross-referenced with PwC, IBM, Deloitte, Gartner, McKinsey, and MIT tells the same story. Luis, unpack it.

LUIS 1:21

The employees are winning, the companies are losing. That's the whole episode.

ELIZABETH 1:26

Give me the evidence.

LUIS 1:30

We analyzed 8,000 AI users in the US across 17 industries. At proficiency, they save 67 minutes per task, four to eight hours a week back in their pocket. And those hours are real.

ELIZABETH 1:46

Our AI for SP Tracker sees the same pattern across 70 countries.

LUIS 1:52

Real and measurable, and almost entirely invisible to the company, because the employee captures the benefit.

ELIZABETH 2:03

Which brings us to the number that should stop every board cold. Eight in ten workers

Shadow AI Shows Real Incentives

ELIZABETH 2:10

now use AI tools the company never approved.

LUIS 2:14

Shadow AI is not rebellion. It's proof of incentive. Every one of those people has a reason. Save time, do better work, go home earlier, think instead of produce all day.

ELIZABETH 2:28

Satisfaction tells the same story. 41% for the tools leadership picked, 78% for the tools people picked themselves.

LUIS 2:38

The individual reason is self-generating. Nobody needs a corporate program to explain it.

ELIZABETH 2:45

So if employees have a clear reason and the productivity is real, why can't companies capture any of it?

LUIS 2:52

Because nobody defined what needed to change. And nobody defined it because the person in charge of defining it does not use AI.

ELIZABETH 3:03

That is the nine in ten number. In nine out of ten failing AI deployments, the leader signing the checks is not a daily user.

LUIS 3:13

You can set direction on something you don't understand. You can build a budget around something you can estimate. You can redesign a workflow when you have never felt what AI does to a workflow. You can

Why Non-User Leaders Lose

LUIS 3:30

name the skills your people need when you don't have those skills yourself.

ELIZABETH 3:35

We called this the Ferrari fallacy in the 56% episode. The car gets bought. The driving lessons get skipped.

LUIS 3:50

And now the driver's seat is empty too. The leader never learned to drive either.

ELIZABETH 3:56

And you can't read your way to the answer.

LUIS 3:59

You cannot. The briefing is not the work. For 30 years, briefings were how leaders did their job. With AI, that habit breaks. You have to use it daily until you understand what your company is actually facing.

ELIZABETH 4:16

Uncomfortable for an executive whose job for years has been to read the briefing, not do the work.

LUIS 4:24

Well, it is uncomfortable and unavoidable.

ELIZABETH 4:28

Before we get to what leaders need to do, I want to come at this from the other side.

LUIS 4:44

Chat GPT had just arrived. And I'll be honest, I was intimidated. You with your background? With my background. And the moment ChatGPT arrived, I felt like I was starting from scratch. Day one, all over again.

ELIZABETH 5:17

What pushed you through?

LUIS 5:19

My dad, he had Alzheimer's, and I was becoming a caregiver for my parents. I needed to keep working, and I needed to carve out two or three hours every single day to be present for them whenever they needed me. There was no plan, there was no path. I didn't even know if it was possible. So I started using it every single day, not because it was strategic, because I had to find hours that didn't exist. And I found them. Artificial intelligence gave me the gift of being

A Personal Reason To Use AI

LUIS 5:55

there for mom and dad in their hardest years. My dad spent his whole life teaching me to figure things out, to do my best with whatever I had in front of me. One more time, without saying a word, he was the one guiding me through this. His legacy is the reason AI for SP exists. It's the reason we now serve 900,000 people in 70 countries. Every person in every company has one. For some people, it's a daughter's recital. For others, it's learning something they've always wanted to learn. For others, it's doing real thinking instead of producing all day. The reason is never the question. The question is whether leadership can see it.

ELIZABETH 6:51

Which takes us to the leaders who do see it. Tell me about the CEO of our largest global consulting client.

LUIS 6:59

When I met with him, he looked me in the eye and said, We are going to grow double digits while transforming for the age of AI.

ELIZABETH 7:12

Still a slogan by itself.

LUIS 7:15

Still a slogan. So we went deeper, and the sentence became we are going to grow double digits without increasing the size of the operation, without burning out our people, and our employee satisfaction is going to go up.

ELIZABETH 7:30

That is the sentence that does the work.

LUIS 8:03

That is the whole difference.

Turning Slogans Into A Real Goal

LUIS 8:05

Not courage, not vision, fluency.

ELIZABETH 8:09

Walk me through how you actually got to that sentence with him, because most leaders never get there.

LUIS 8:16

We did a listening tour first. 7,000 people. Every level, every geography. What are you using? What are you saving? What's blocking you? By the end, we had a map of where hours were already being won inside the firm and where the workloads were fighting the tools.

ELIZABETH 8:35

So the sentence came from evidence, not from a whiteboard.

LUIS 8:39

From evidence. And then we changed one thing that mattered. We restructure how delivery teams got paid. When a team produced more with the same headcount, they share the app site directly.

ELIZABETH 8:56

One mechanism, not a framework.

LUIS 9:00

One mechanism, because frameworks don't change behavior. Rewards do. That one change told every person in the firm the same thing. The company's reason and your reason are now the same reason.

ELIZABETH 9:15

Now that is a retrofit, a non-native company rebuilding compensation around what AI does. Contrast that with companies that were built that way from day one. Anthropic crossed $30 billion in revenue with about 5,000 people. Cursor hit $2 billion in three years with a few hundred employees.

LUIS 9:40

And everyone assumes it's the tools. It isn't. Everyone has the same tools. The difference is that the founders are daily users of AI. They build the company knowing what AI can do. The

Incentives That Capture Productivity

LUIS 9:55

strategy, the org design, the compensation, the workflows, all of it lines up with what the technology actually does.

ELIZABETH 10:04

Elizabeth, well, do the math per employee. Anthropic is around $6 million of revenue per head. Cursor is in the same range or higher.

LUIS 10:15

And before anyone writes this off as tech companies are just more efficient, Apple, Google, Microsoft, Meta are in the 1 million to 2.5 million range. Anthropic is running two to five times ahead of the most efficient pre-AI companies in the world.

ELIZABETH 10:35

So the company's reason and the employees' reasons are aligned by design.

LUIS 10:40

By design, not retrofitted. Every person uses AI daily. Every role is a structure around that. Every incentive rewards that. Non-native enterprises try to do the opposite. They announce a transformation, but the OR chart is inherited. Compensation rewards yesterday's work, and the workflows stay the same. The company's reason and the employees' reasons never meet.

ELIZABETH 11:07

And the use of tools not approved by the company, also called shadow AI, is what fills the gap.

LUIS 11:14

Shadow AI is the gap invisible form.

ELIZABETH 11:18

So how does a leader who was not born AI native actually close it?

LUIS 11:24

Two things. Exactly. Shadow AI,

AI Native Design Versus Retrofit

LUIS 11:40

or the use of tools not authorized and paid for by the company. Eight in ten of your people are using unauthorized tools to run real experiments every day with real workloads. They know which tools work. They know where AI breaks. They know the friction. That is not a compliance problem. That is a learning lab.

ELIZABETH 12:03

And most leaders never open the door.

LUIS 12:06

They shut it, they send a policy memo, they block the tools. They miss the best piece of intelligence their company has ever had.

ELIZABETH 12:16

In fairness to those IT and compliance teams,

Treat Shadow AI Like A Lab

ELIZABETH 12:20

they are following the rules they were given. The rules are what need to change.

LUIS 12:26

Okay, fair point. Change the rules, change the posture. And while you're changing them, listen.

ELIZABETH 12:34

This is exactly why we always start with diagnosis and listening to your people. Not just to leaders, not just to a small group, to every single one. What they use, what they save, what they struggle with, surfaced so leaders can actually act on it.

LUIS 12:53

Daily use teaches you what AI can do in the abstract. Listening teaches you what AI is already doing inside your company. Put the two together and you have what you need to set real strategy.

ELIZABETH 13:06

And only then does change management work, because only then do leaders know what to change.

LUIS 13:13

Exactly. You can manage a change you don't understand. Change management without fluency is just motion. Deck chairs get shuffled, the ship goes nowhere.

ELIZABETH 13:24

So what is the one takeaway for our leaders today? Three questions.

LUIS 13:30

First, are you using AI daily enough to know what it can actually do in your company? Second, are you listening to the people in your company who are already using it? Third, are you willing to change how you manage, how you measure, how you reward, and how you organize work so the company's reason and your people's reason

Three Questions For Real Transformation

LUIS 13:55

finally meet? If you can't answer those three, you don't have an AI transformation. You have a press release

ELIZABETH 14:03

All sources and companion article are at A.I. for S.P. dot org. To learn more, ask your favorite A.I. assistant about us[short pause] Stay curious, and be kind to each other.