Private Equity Firm Bets AI Will Eliminate Corporate Departments

Small companies positioned to lead AI transformation

A private equity operating partner is making an unconventional prediction: traditional corporate departments will cease to exist within five years, replaced by AI systems that cut across organizational boundaries.

Bob Root, operating partner at Southfield Capital, argues that finance, purchasing, and marketing will no longer function as separate departments. His firm is already building custom AI solutions for portfolio companies that ignore conventional departmental structures entirely.

Southfield focuses on founder-run businesses in the lower middle market—companies generating roughly $5 million to $10 million in EBITDA. Root's thesis holds that these smaller firms, historically underserved by technology, now possess a critical advantage: minimal approval layers and faster decision-making compared to large corporations.

Why it matters

While most private equity firms discuss AI in terms of cost reduction and margin expansion, Southfield's approach represents a fundamental reimagining of organizational structure. The firm's bet challenges decades of corporate hierarchy and raises questions about whether AI will truly obliterate departments or simply lead to incremental adjustments—a debate that echoes failed reengineering promises from the 1990s.

Building around proprietary data, not departments

Southfield's portfolio spans unglamorous sectors: industrial refrigeration, commercial landscaping, property management, and security services. In one field service business, the firm built an accounting and procurement tool with portfolio company Contextual.io that handles pricing, purchase orders, and approvals automatically—designed without regard to traditional departmental boundaries.

Root describes AI as a "fourth leg" that cuts across people, process, and technology, functioning more like an operating system than a tool. He dismisses off-the-shelf solutions as commoditizing rather than differentiating, instead pursuing what he calls a "data moat thesis"—identifying proprietary knowledge during due diligence that competitors lack.

This approach contradicts broader industry data. An MIT Project NANDA preliminary report found approximately 95% of enterprise generative AI pilots produced no measurable profit and loss impact. The same research showed externally partnered tools reached deployment roughly twice as often as internally built ones.

CEO ownership as prerequisite

Root enforces a strict rule: "If it's not owned by the CEO, we don't do it." He describes clients as living with "some sort of level of AI paranoia"—knowing something must be done without knowing what.

To address this, Southfield facilitates peer conversations between portfolio CEOs and runs one-day workshops where teams reimagine operations from scratch. Even then, discipline proves challenging. Root recalls an industrial refrigeration CEO who, after initial skepticism, attempted to expand a focused two-project initiative into eight simultaneous ones, diluting returns.

The broader AI transformation landscape remains sobering. S&P Global Market Intelligence's 451 Research survey found the proportion of organizations abandoning most AI initiatives jumped from 17% to 42% in one year, with the average organization scrapping 46% of proofs of concept before production.

Root's prediction echoes Michael Hammer's 1990 Harvard Business Review call to "obliterate" rather than automate, replacing functional silos with cross-functional processes. Most organizational charts survived that wave. Whether AI proves different remains an open question.

These details were first reported by Charles Towers-Clark for Forbes.