A new study from PwC is putting a precise number on what many business leaders already sense: AI’s economic gains are not being distributed evenly.

The 2026 AI Performance Study found that just 20% of companies are capturing approximately 75% of AI’s total economic value. The other 80% — implementing AI in some capacity, running pilots, adding tools — are largely leaving gains on the table.

The gap isn’t about access to technology. It’s about how companies are using it.

Leaders vs. Followers: What’s Actually Different

PwC’s research identifies a clear behavioral divide between the top performers and the rest.

AI leaders are using AI to:

• Pursue new revenue streams and business models
• Redesign core workflows from the ground up
• Make AI a strategic priority at the executive level

AI followers are using AI to:

• Automate existing tasks incrementally
• Reduce headcount or costs at the margins
• Add AI features to existing products without rethinking the product

The study found that leading companies are 2–3x more likely to redesign workflows around AI rather than simply layering AI onto current processes.

Why Incremental Adoption Is Falling Behind

This is the uncomfortable finding for most organizations: doing something with AI is no longer enough.

When AI is treated as a productivity tool — a way to do the same things faster — companies tend to see modest, hard-to-measure gains. The technology improves individual tasks but doesn’t change the underlying business model or competitive position.

When AI is treated as a platform for reinvention, the results compound. New workflows create new capabilities. New capabilities enable new offerings. New offerings attract new customers. That flywheel is what separates a 2x outcome from a 10x one.

The window for making that shift is narrowing. As AI infrastructure becomes commoditized, the competitive advantage will come from organizational design, data assets, and execution — not from access to the tools themselves.

Three Areas Where Leaders Are Pulling Ahead

1. Customer Experience Reinvention

AI leaders aren’t just using chatbots for support — they’re redesigning the entire customer journey. Personalization at scale, real-time pricing, dynamic product recommendations, and proactive outreach are happening automatically, informed by behavioral data and AI inference. Customers expect this. Companies that can’t deliver it will lose ground to those that can.

2. Operations and Supply Chain

Top performers are using AI to make operational decisions — demand forecasting, supplier risk assessment, logistics optimization — that previously required large teams of analysts. The cycle time from data to decision is collapsing. What took weeks now takes hours. That speed advantage compounds across hundreds of decisions per year.

3. New Business Model Creation

This is the highest-leverage use of AI, and the most underused. Companies like insurance carriers, financial services firms, and retailers are discovering entirely new revenue streams by productizing their AI capabilities — offering intelligence-as-a-service to partners and customers. This was not possible at scale without AI.

What the Laggards Are Getting Wrong

The PwC study doesn’t frame the gap as a technology problem. It frames it as a strategy and ambition problem.

Companies in the bottom 80% tend to:

• Run disconnected AI pilots with no path to scale
• Measure success by cost savings rather than value creation
• Lack executive alignment on AI’s strategic role
• Underinvest in data infrastructure and AI talent

None of these are technical barriers. They are organizational ones.

What to Do About It

If your organization falls into the follower category — and statistically, most do — here are the moves that PwC’s research suggests have the highest leverage:

Start with a business model question, not a tool question. Instead of asking “where can we use AI?”, ask “what would our business look like if AI made our biggest bottleneck disappear?” Then work backwards.

Pick one high-stakes workflow to redesign completely. Not incrementally improve — completely rethink. What assumptions does the current process make that AI invalidates? Build around the new assumptions.

Measure value creation, not cost reduction. AI projects that target revenue, new customers, or new products get different executive attention — and different resource allocation — than projects that target headcount reduction.

Build for compounding. The goal isn’t to extract a one-time benefit from a single AI deployment. It’s to build systems that get better over time as they ingest more data and generate more feedback loops.

The Bottom Line

PwC’s findings confirm what’s becoming visible in market outcomes: the AI divide is real, measurable, and widening. Three-quarters of the value is going to one-fifth of participants — not because those companies have better technology, but because they made a deliberate choice to use AI to change how they compete, not just how they operate.

The technology is available to everyone. The ambition and organizational will to use it strategically is not. That’s the actual differentiator right now.

For business leaders, this study is a useful diagnostic. If your current AI strategy is primarily about doing existing things more efficiently, you’re likely in the 80%. The question is whether you want to stay there.