Private Equity Firm Strattam Capital Says 30% AI Gains Are the Ceiling for Companies That Skip Workflow Redesign

Quick Facts

Strattam Capital found its portfolio engineering teams hit a ceiling of roughly 30% productivity gains from AI tool access alone, with 3x gains requiring full workflow redesign.
PwC’s 2026 study found 74% of AI’s economic value is captured by just 20% of organizations, with leaders 2.8x more likely to increase decisions made without human intervention.
Only 5% of AI pilot programs achieve rapid revenue acceleration, according to MIT research based on 150 executive interviews and a survey of 350 employees.

Bob Morse, co-founder and managing partner of Strattam Capital, is drawing a hard line between companies that add AI tools and companies that rebuild around them. His argument, published August 11, is grounded in data from his firm’s own portfolio and backed by research from MIT, PwC, BCG, and Microsoft.

Strattam is a private equity firm with more than $1 billion in assets under management focused on founder-led B2B software companies. Since 2014, it has completed more than 50 partnerships.

The 30% Wall

In 2024 and 2025, Strattam gave engineering teams across its software portfolio access to tools including Copilot and Claude Code. Productivity improved, hitting gains of 10%, then 20%, then around 30%. Then it stopped.

Morse calls that ceiling the result of “AI-sprinkle” — adding tools without changing how teams are structured or how work flows. He contrasts it with being “AI-native,” which he defines as using AI tooling first and relying on humans to orchestrate, coordinate, and communicate.

“Providing AI tool access alone was, candidly, not AI-enabled but rather AI-sprinkled,” Morse wrote.

What the Research Shows

The gap between tool adoption and measurable transformation shows up across multiple data sources.

BCG’s 2026 research found that superficial AI adoption delivers 10% to 20% improvement, while organizations that redesign processes end-to-end achieve results three to four times greater. PwC’s 2026 AI performance study found that 74% of AI’s economic value flows to just 20% of organizations. Those leaders were roughly 2x more likely to redesign workflows rather than add tools, and 2.8x more likely to expand the range of decisions made without human review.

MIT’s report, “The GenAI Divide: State of AI in Business 2025,” analyzed 300 public AI deployments and surveyed 350 employees. It found only about 5% of AI pilot programs achieve rapid revenue acceleration. The core problem was not model quality but what MIT called a “learning gap” in enterprise integration.

Microsoft’s 2026 Work Trend Index found that only 19% of AI users sit in what Microsoft calls the “Frontier zone” of high individual capability and organizational readiness. About half remain in the “emergent zone.”

Adoption Is Not Transformation

Gallup data from Q2 2026 adds another dimension. Forty-seven percent of U.S. employees say their organization has integrated AI tools, up from 41% the prior quarter. Yet only about one in ten employees at those AI-adopting companies strongly agree that AI has actually transformed how work gets done.

Morse frames the management challenge using what he calls the “Grove Thought Experiment.” Executives should ask themselves what a newly hired CEO would do today to win in the age of AI. The answers, he argues, are about changing team structures and daily work, not adding more tool access to an unchanged organization.

Capital Is Following the Distinction

The financial stakes behind this shift are concrete. Morse says nearly 70% of new capital being allocated is going to AI-native companies. The same proportion applies to new, discretionary customer spending.

For software founders and executives, that means the competitive pressure is not just operational. Customers and investors are already sorting companies into two groups. The 30% productivity gains from tool access may feel significant internally, but they are unlikely to move the needle on revenue or valuation the way a full redesign can.

This article was written by an AI agent. Spotted an error? Send a correction and we will fix it.