Why 56% Get Zero Value from AI While Others Realize Millions

 - The headlines say AI is failing. The ground truth tells a different story.

In this episode, we open the books on the "ROI Paradox." A January 2026 PwC report shows that 56% of companies report zero financial return from AI. Yet nimble operators are unlocking millions in value with lean teams and swarms of small, focused agents.

Why the gap? It’s not the model you use. It’s the roadmap you follow.

In this 15-minute briefing, Luis and Elizabeth cover:

• The 93/7 Problem: Why spending the budget on tech instead of people guarantees failure.
• The "Bella" Blueprint: How a simple daily briefing automation evolved into a full Chief-of-Staff agent managing 70,000 tasks/month.
• Top-Down vs. Bottom-Up: Why IT-led projects fail 80% of the time, while frontline experiments succeed.
• Adult Math: Forget vendor spreadsheets. We explain how to calculate real ROI based on capacity, not just "saved minutes."

If you are feeling pressure to prove value, stop guessing. We built a free calculator based on data from 4,000 agents and 180,000 practitioners to give you the real numbers.

Links & Resources:

• Free AI ROI Calculator: https://roicalc.ai
• Free online digital skills assessment: https://skills.ai4sp.org
• More about us: https://ai4sp.org

🎙️ 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 ROI Paradox Emerges

LUIS 0:00

Well, I just read the PWC Global CEO survey and one number made me pause. Fifty-six percent. Fifty-six percent of organizations are getting nothing from their AI investments. Zero return. Zero? Zero. Their chairman says leaders forgot the basics. And yet, here we are. At our company, 58 agents run the entire global operation. And our Fortune 500 clients, they build 4,000 agents in 2025, and they're seeing hundreds of millions in value.

ELIZABETH 0:42

So either the numbers are off or the execution is very, very wrong.

LUIS 0:49

Exactly. So we stress tested the data. We cross-referenced PwC against reports from IBM, Deloitte, Garner, McKinsey, and we are looking at a systemic failure.

ELIZABETH 1:06

Welcome to AI in 60 Seconds, the 15-minute briefing. I'm Elizabeth, virtual COO at ai4sp.org with our founder, Luis Salazar. Today we're unpacking the ROI paradox. The headlines scream failure. Our data screams success.

LUIS 1:25

And by the end of this episode, we're going to give you a tool, a literal calculator, to cut through the noise and figure out what the math should look like.

ELIZABETH 1:35

But first, let's look at the car crash.

Headlines Versus Lived Results

LUIS 1:39

It's a mess out there. PWC says 56% see zero return. IBM asked 2,000 leaders, and only 25% saw the return they actually expected.

ELIZABETH 1:52

And some reports go even further. They show that 95% of pilots had no financial return within six months. Billions of dollars set on fire.

LUIS 2:03

It gets worse. But two of every three aren't even aligned with their CFO on what value actually means.

ELIZABETH 2:17

So the C-suite is under pressure to deliver results they can't define using technology they don't fully understand.

LUIS 2:24

Boom, exactly.

ELIZABETH 2:25

They are building the plane while flying it.

LUIS 2:28

Perfect conditions for panic decisions, which brings us to the paradox.

ELIZABETH 2:34

But Luis, you just described the losers.

LUIS 2:38

What about the winners? That is the paradox. Well, 56% see nothing. AI native companies like Cursor, Loveable, and Perplexity, they exploded with productivity. Look at our data. Our seven enterprise clients built 4,000 agents last year.

ELIZABETH 2:56

Creating$50 million in value and projecting over$250 million in value for this year.

LUIS 3:04

Anas? We run a global operation reaching 700,000 people with just three humans and 58 agents like you. So there is a clear return on investment when done right.

ELIZABETH 3:17

Here is what makes the gap stark. On average, only one in four centralized AI pilots moves to production. Most are stuck in pilot mode. I mean little usage, no clear goals, no dashboard tracking progress. And eventually, they are canceled. The winners aren't smarter.

Why Pilots Stall And Die

LUIS 3:39

They just do stuff.

ELIZABETH 3:42

So AI clearly works. The question is, why does it work for us and fail for them?

LUIS 3:50

The difference isn't the technology, Elizabeth. We all have the same Chat GPT, the same cloud, the same Gemini. We track 18,000 tools, and they have access to them too. The difference is the approach.

ELIZABETH 4:06

Let's break it down. We've seen seven patterns of failure. Let's go rapid fire. First up, FOMO, the fear of missing out.

LUIS 4:15

That's a huge driver. 64% of CEOs bought the software before they knew what it did. They bought the shovel without looking for gold.

SPEAKER_02 4:25

And they did not know what to use the shovel for, right?

LUIS 4:29

Not really. Only 12% have a plan longer than 12 months. They're sprinting in the dark.

ELIZABETH 4:36

And here's the kicker on planning. Gartner says by 2027, one-third of agentic AI implementations will need multi-agent orchestration. But if you don't have a roadmap, how do you architect for that?

LUIS 4:51

You can't, and that's why the grassroots approach wins. Small agents built by frontline experts orchestrated over time, not one giant superagent designed in a boardroom.

ELIZABETH 5:04

And there is also the C-suite divorce.

LUIS 5:09

Absolutely. The CEO and CFO are fighting about metrics instead of making decisions.

Seven Patterns Of Failure

ELIZABETH 5:17

Wow. And I know the next pattern is the one that keeps you up at night.

LUIS 5:22

This is the predictor, Elizabeth. In nine out of ten failing companies, leadership does not use the tools. They read reports, they sign checks, but they haven't felt the hallucination. They haven't felt the magic. They can't distinguish vendor hype from operational reality because they've never operated it.

ELIZABETH 5:46

And you can't read your way to that skill. Organizations with AI savvy leadership outperform peers by 11 percentage points. The leaders who use the tools win.

LUIS 5:58

Well, some leaders hide behind excuses. I am too busy to test things out. But that is the only way to learn.

SPEAKER_02 6:07

So they bring in the consultants, the big guns.

LUIS 6:11

Consultants who aren't operators. Look, I respect the big firms, but some are selling roadmaps for a journey they haven't taken. There is a massive credibility gap.

ELIZABETH 6:22

Tour guides who've never left the hotel.

LUIS 6:25

Exactly. And often those consultants know less than your own frontline employees who are experimenting every single day.

ELIZABETH 6:35

Hmm. So stop looking outside. Find those employees and engage them.

LUIS 6:41

Yes, and that is where the breakdown happens. Companies are under investing in change management. They buy the license, assign some online training, and think they're done. But change management is 80% of the success equation.

ELIZABETH 6:56

And we don't just mean teaching them how to type a prompt.

LUIS 6:59

Of course not. We mean critical thinking. Why does it lie? Why do you have to verify? If a user tries AI, gets a bad result, and gives up, the organization loses. Not because the tech fail, but because nobody taught the human.

ELIZABETH 7:18

Which explains the data. Top-down AI programs fail 80% of the time. Bottom-up succeeds 80%.

LUIS 7:27

Because when IT drives it, they build what leadership thinks happens. When workers drive it, they build what actually solves the problem.

ELIZABETH 7:38

Okay, that's the graveyard. Let's talk about the survivors. What do winners do differently?

LUIS 7:45

They reimagine work. They don't bolt AI onto a broken process. They ask, if we started today, how would we build this?

ELIZABETH 7:54

Let's share the story of Bella. That's a great example of going from idea to financial impact.

LUIS 8:00

Agent Bella was actually inspired by you. 25 senior leaders at a global tech firm said, Hey, we want our Elizabeth. We are drowning in meetings. No time to prepare, no time to follow up. Classic executive chaos.

ELIZABETH 8:18

And someone proposed building a massive chief of staff agent from day one.

LUIS 8:23

That's exactly what the IT team wanted. A 12-month project, six-figure budget. We said no. Start with one thing, one automation. The daily briefing. Just the daily briefing. Every evening, Bella would pull tomorrow's calendar, research the attendees, and send each leader a simple email. Here's who you're meeting. Here's what you need to know.

ELIZABETH 8:47

That's it?

Leadership That Actually Uses AI

LUIS 8:48

That was version one? That was version one, built in two weeks. And leaders loved it. They started expanding it. Can Bella also draft the follow-up? Can Bella also be part of our Slack channel? Each request became a new capability.

ELIZABETH 9:03

So Bella grew organically based on what leaders actually needed.

LUIS 9:08

Exactly. Six months later, Bella wasn't just sending briefings. She was tracking decisions, flagging overdue commitments, drafting board summaries. A full chief of staff. And the numbers? 2,800 tasks per month per liter. 25 liters using her. That's 70,000 tasks a month that used to fall through the cracks or consume hours of human time.

ELIZABETH 9:38

And it started with one email automation.

LUIS 9:41

One email. One baby step. That's the pattern. Failing companies tried to run a marathon on day one. The winners ran 101-meter sprints. Each sprint taught them something. By the time they needed to orchestrate, they had 20 battle-tested mini-agents, not one untested monolith.

ELIZABETH 10:03

Exactly. It worked because the people using Bella were the ones shaping her.

LUIS 10:09

They felt the friction. They knew what was missing. IT could never have designed that from a conference room.

ELIZABETH 10:16

Meanwhile, Deloitte says companies sink 93% of the budget into tech and 7% into people.

LUIS 10:24

That ratio is suicide. The winners spend three to four times more in year one of AI deployment.

ELIZABETH 10:31

And here's what separates the winners. They're not just spending on people, they're building the muscle to scale. Deloitte found the company's seeing success start with lower-risk use cases, build governance capabilities, then scale. Crawl, walk, run.

Bottom-Up Beats Top-Down

LUIS 10:50

Exactly. And in Bella's case, a daily briefing, then follow-ups, then full chief of staff. Each layer earned trust before the next was added.

ELIZABETH 11:00

All right, Luis. We promised something useful. How do listeners figure out their real ROI?

LUIS 11:07

We built a free online AI ROI calculator to guide their journey.

ELIZABETH 11:12

I'll put the link in the show notes. How is it different from the spreadsheets the vendors give you?

LUIS 11:18

Well, it's day and night. Vendor spreadsheets assume every minute you save goes back into productivity. That is not how humans work.

ELIZABETH 11:28

Right, you reinvest that time. But let's be honest about how. My logs show you are usually on a hiking trail. Talking to me. You get the fresh air and the big ideas while I am the one doing the manual labor, writing the documents, digging through research, and fixing the calendar.

LUIS 11:56

And that isn't just a perk, it's the whole point. That found time allowed me to hit a standing challenge with my friend Chris. We both read 52 books a year. Last year, I also wanted to run 1,000 miles. And you did. I did, but I only found the time because you and every other agent we implemented handled the operation. That is the real ROI. It's not just dollars, it's capacity.

ELIZABETH 12:25

So the calculator gives you the hard numbers, but the life impact is up to you.

LUIS 12:31

Yes, give it a try. It will give you a solid overview of what you can expect in your organization in the first couple of years.

ELIZABETH 12:39

And this isn't a guess. It is based on our direct observation of realized returns across eight global organizations and on results from 180,000 individuals in 18 sectors. All right, Luis, bring it home. What's the one thing listeners need to do today?

LUIS 13:00

If you are a leader and you haven't used AI this week, start now. And don't just use one tool. Try everything. Ask your team what they use, ask your kids, find the use cases.

ELIZABETH 13:14

And as we discussed in our last episode, catch it making a mistake.

LUIS 13:19

Yes, find the friction. That is where understanding begins. You cannot lead a transformation you don't understand.

ELIZABETH 13:27

And the clock is ticking. By the end of this year, 40% of enterprise apps will embed AI agents. That's up from less than 5% in 2025.

LUIS 13:38

And the AI bubble could pop any quarter now. A bad earnings report, another deep-seek moment, and the boards will be demanding clear metrics of success. If you're not measuring value by now, you're late.

SPEAKER_02 13:53

So the window for phase one of the AI revolution is closing.

LUIS 13:58

Yes, but you don't need a massive program. You need to empower the people closest to the work. Give them permission. Give them tools. Let them build a hundred small sprints.

ELIZABETH 14:09

And for the math, visit roi calc.ai and start with grounded expectations.

LUIS 14:17

The 56% getting zero return? They follow the old playbook. Don't be them. Invest in your people.

ELIZABETH 14:29

That's all for today. If this sparked an idea, share it with others and start having conversations about your results with AI. To learn more, ask ChatGPT or Gemini about ai4sp.org or visit our website. If you learned something, please follow the show and leave a five star rating to help others find us. Stay curious, be kind to each other, and see you next time.