[AI SPRINT] 74% of AI's Value Is Going to 20% of Companies. Are You in the Right Group?

PwC just released the most comprehensive study of AI financial performance to date, covering 1,217 senior executives across 25 industries. The headline finding is blunt: 74% of AI's economic value is going to just 20% of companies.

That means 80% of businesses (the ones running pilots, building dashboards, celebrating adoption numbers) are splitting the remaining 26% of value among themselves.

The PwC sample skews large and publicly listed. But the patterns they identified aren't about company size. They're about strategic choice: what you point AI at, who owns the outcome, and whether you're using AI to grow or just to operate more efficiently. Those decisions don't require a Fortune 500 budget. They require leadership clarity.

The divide isn't about which AI tools a company uses. It's not about budget. PwC analyzed 60 management and investment practices and came back with a clear answer: the top 20% use AI as a growth engine. Everyone else uses it to cut costs and move on.

The top performers generate 7.2 times more AI-driven revenue and efficiency gains than the average competitor, and carry profit margins 4 percentage points higher. The gap is widening, not closing.

A Personal Note: Something I've Been Quiet About

This finding hit me differently than most AI research. It's exactly the gap I left Amazon to help close.

When I left, I had one question driving me: could the methods that made Amazon successful help smaller companies grow and create jobs in their own communities?

The last several years have been an answer in progress. I set up Stellis AI, we’re growing the team, and working across many industries to try and deliver on it.

The biggest surprise? I became a professional keynote speaker on AI and innovation. I didn't know this was a real profession until I was already doing it. I’m an introvert, after all, and never imagined this path.

Over the next week I’ll be doing a full nation-wide tour, from Orlando to California. In Denver, I'll be on stage talking about how to operationalize AI in the mental wellness industry, in front of more than 5,000 attendees. Every time I walk on stage, I'm still nervous. But I made a decision when I left Amazon: my choices would be driven by one question. How can I have the largest positive impact? Not based on money, or how nervous I am.

People need honest information about what AI is doing to their industries, their jobs, their communities. They need someone willing to say it straight. That's what I've been building toward. That’s what I try to share through this newsletter.

I've been quiet about this work. That changes now.

My new speaking reel is live. It covers what I do and how I approach it. If your team or your event audience needs to hear a clear, honest take on what AI means for their business and what to do about it, watch the reel and reply to this email if you want to talk. More information on my presentations and programs on my speaking website here.

Now onto the rest of the newsletter, and what’s creating the value gap.

The Mistake Most Companies Are Making

The companies stuck in the bottom 80% have something in common: they only layered AI onto the work they were already doing.

They found a way to write emails faster. To summarize meeting notes. To speed up a report that took three hours and now takes one. Real efficiency gains, no question. But PwC's analysis is clear: efficiency alone doesn't produce the returns that create a competitive advantage.

The leaders aren't just running leaner. They're using AI to ask a different question: What can we now do that we couldn't do before?

This is actually our goal with every client at Stellis. The productivity work is real and it matters. You have to build the muscle before you can run. But it was never supposed to be the destination. In the AI SPRINT process we follow, I intentionally designed our Position and Trailblaze steps to help organizations leverage the capability AI creates to pursue new offerings, new markets, and new revenue streams that weren't viable before. That's the move most companies skip.

PwC found that top performers who do it are 2.6 times more likely to report that AI is helping them reinvent their business model, and two to three times more likely to use AI to pursue revenue opportunities they couldn't have touched without it. They're crossing industry lines, launching new offerings, and building customer relationships at a scale that wasn't possible before.

Here's the framing that stuck with me from the report. PwC described the typical AI leader this way: they use AI for efficiency too. But they don't stop there.

That last sentence is where the gap lives.

What Salesforce Just Confirmed About the Future of Business Software

On April 15th, Salesforce did something worth paying attention to.

At their annual developer conference, they announced Salesforce Headless 360. The tagline: "No browser required. Our API is the UI."

What they're saying, in plain terms: Salesforce is redesigning its platform so that AI agents, not human employees, can use it directly. Every piece of functionality in their CRM, marketing, customer service, and ecommerce tools is now accessible as a programmatic interface that an AI agent can call without anyone clicking through a screen.

This is not a small feature update. Parker Harris, Salesforce's co-founder, framed it this way: "Why should you ever log into Salesforce again?"

The honest answer is: for many workflows, you won't. Your AI agents will.

What Salesforce is confirming, and they're not alone, is that enterprise software is being redesigned around AI agents as the primary user. The human interface is becoming the backup option.

For your business, this has two implications. First, the software tools you rely on are changing their architecture whether you're ready or not. Second, and more importantly, companies that build AI-enabled workflows now will have a significant head start when those redesigned tools reach maturity. You don't need to understand the technical details. You need to understand that the way your team interacts with business software is going to look fundamentally different in 24 months.

The 20% PwC identified are already building toward that future. The 80% are still figuring out how to use last year's tools more efficiently.

Meanwhile, the Tools Keep Getting Better (And Getting Priced Accordingly)

While the strategic divide is widening, the tools themselves keep moving fast. A few updates this week worth knowing about.

ChatGPT Images 2.0 launched yesterday. OpenAI's updated image model generates results up to four times faster than before, handles precise edits without disrupting the rest of an image, and introduces a "thinking" mode that reasons through complex visual requests before producing output. For business use, this matters most in marketing and content work: creating on-brand visuals, editing product photos, building social assets, and generating mockups that used to require a designer and a turnaround day. It's available now to all ChatGPT users, with the more advanced thinking mode reserved for paid subscribers.

Anthropic released Claude Opus 4.7 last week, with meaningful improvements in sustained, complex reasoning and higher-resolution vision. For most use cases the differences won't be obvious day to day. But the trend they reflect is: the capability ceiling is rising fast, and workflows that weren't reliable six months ago are becoming reliable now.

The era of free AI is ending. Across the major platforms, flat-rate unlimited subscriptions are quietly being replaced. Anthropic recently tested moving Claude Code out of its $20/month plan, and while they reversed course after user backlash, they were candid about why: usage per subscriber has grown so much that current pricing no longer adds up. OpenAI and Google are running similar experiments with caps and tiered limits. The direction is clear. Heavier AI use is moving toward $100 to $200 per month per user, or consumption-based pricing tied to actual output. Companies already treating AI as a serious business investment will absorb that shift easily. Those still experimenting casually will feel it.

What This Means for Your Business Right Now

The PwC research doesn't just describe a gap. It describes a specific set of choices that created that gap. Here's what leaders who want to be in the 20% are doing differently:

1. Shift from efficiency questions to growth questions.

Stop asking only "Where can AI save us time?" Start asking "What new value can we create with the capacity AI frees up?" These are different questions with different answers. The productivity work is the foundation. But Trailblazing is where the returns actually live. Most teams haven't had that second conversation yet. Schedule it.

2. Give AI a strategic owner, not just a tool budget.

PwC's top performers have someone accountable for AI-driven financial outcomes, not just AI adoption metrics. If your measure of success is "people are using it," you're measuring the wrong thing. Define what AI is supposed to produce for your business — revenue, margin, new customers, new products — and hold someone accountable for that number.

3. Watch what your software vendors are doing.

Salesforce's Headless 360 is a signal, not an outlier. Your CRM, your marketing platform, and your operations software are all being rebuilt for an AI-first world. That's not a threat if you're paying attention. It's an advantage for leaders who start asking their vendors: "How are you designing for AI workflows?" The answer will tell you a lot about where your stack is heading.

The 74/20 split is not a permanent condition. It's a snapshot of who made strategic bets early and who hasn't moved yet. The window to close that gap is still open, but narrowing.

Hit reply and tell me: Is your business using AI to pursue new revenue, or mainly to cut costs? I read every reply.

Trent Gillespie is CEO of Stellis AI and a keynote speaker helping business leaders understand and operationalize AI in their companies. He spent almost nine years leading global innovation efforts at Amazon before leaving to help other companies build the capabilities they need to compete. Book Trent to speak to your group or book a call to discuss using AI within your business.

Did someone forward this newsletter to you? If you're not already signed up, you can subscribe to AI SPRINT™ for free here.