Venture capitalists are increasingly adopting a new investment strategy, aiming to leverage artificial intelligence (AI) to extract software-like margins from traditionally labor-intensive services businesses. This approach involves acquiring mature professional services firms, integrating AI to automate tasks, and then utilizing the enhanced cash flow to facilitate further company acquisitions in a "roll-up" model.

General Catalyst (GC) stands at the forefront of this trend, having allocated $1.5 billion from its latest fundraise to what it terms a "creation" strategy. This strategy focuses on incubating AI-native software companies within specific industry verticals, which then serve as acquisition vehicles to purchase established firms and their customer bases in those same sectors. GC has already initiated investments across seven industries, including legal services and IT management, with ambitious plans to expand into up to 20 sectors. Marc Bhargava, who spearheads GC’s related efforts, highlighted the significant disparity between the $16 trillion global services market and the $1 trillion software market, noting that software's appeal lies in its inherently higher margins due to minimal marginal costs at scale. He posits that by automating 30% to 50% of tasks in services businesses with AI, and even up to 70% in high-volume areas like call centers, the financial prospects become compelling.

The efficacy of this game plan is already evident in GC's portfolio. For instance, Titan MSP, a GC portfolio company, received $74 million across two tranches to develop AI tools for managed service providers. Following this, it acquired RFA, a prominent IT services firm. Pilot programs demonstrated Titan's capacity to automate 38% of typical MSP tasks. The company now intends to capitalize on its improved margins to acquire additional MSPs, executing a classic roll-up strategy. Similarly, GC incubated Eudia, a firm specializing in in-house legal departments, which has secured Fortune 100 clients such as Chevron, Southwest Airlines, and Stripe. Eudia delivers fixed-fee legal services powered by AI, diverging from traditional hourly billing. Its recent acquisition of Johnson Hanna, an alternative legal service provider, further broadened its market reach. Bhargava indicates GC's objective is to at least double the EBITDA margin of the companies it acquires.

This strategic thinking extends beyond General Catalyst. Venture firm Mayfield has committed $100 million to "AI teammates" investments. One such investment, Gruve, an IT consulting startup, acquired a $5 million security consulting company and subsequently escalated its revenue to $15 million within six months, while achieving an impressive 80% gross margin, according to its founders. Navin Chaddha, Mayfield’s managing director, remarked that if 80% of the work can be performed by AI, gross margins could reach 80% to 90%, potentially yielding blended margins of 60% to 70% and net income of 20% to 30%. Independent investor Elad Gil has also pursued a similar strategy for three years, backing companies that acquire mature businesses and transform them with AI, arguing that direct asset ownership allows for more rapid transformation than merely selling software as a vendor.

However, emerging data suggests this AI-driven services industry transformation might be more complex than venture capitalists anticipate. A recent study by researchers at Stanford Social Media Lab and BetterUp Labs, surveying 1,150 full-time employees, revealed that 40% of respondents experienced an increase in workload due to "workslop" – AI-generated output that appears polished but lacks substantive quality, consequently creating additional work and challenges for colleagues. Employees reported spending an average of nearly two hours dealing with each instance of workslop, including deciphering, deciding whether to return it, or fixing it themselves. The study estimated that workslop imposes an "invisible tax" of $186 per month per person, translating to over $9 million annually in lost productivity for an organization of 10,000 workers.

Despite these findings, Marc Bhargava of GC disputes the notion that AI is overhyped, instead interpreting implementation failures as validation of GC’s approach. He contends that the difficulty in effectively applying AI technology to businesses underscores the opportunity. If large corporations could easily deploy AI through consulting firms or simple software contracts, GC's thesis would be less robust. Bhargava emphasizes the critical need for technical sophistication, particularly "applied AI engineers" from leading tech companies, who possess a deep understanding of various AI models, their nuances, and optimal application. He argues that GC's strategy of combining AI specialists with industry experts to build companies from the ground up directly addresses this complexity.

Nonetheless, workslop poses a potential threat to the core economics of this strategy. Should acquired companies reduce staff based on projected AI efficiencies, they would have fewer personnel to identify and rectify AI-generated errors. Conversely, if staffing levels are maintained to manage the additional work created by problematic AI output, the substantial margin gains that VCs expect might not materialize. These scenarios could necessitate a slowdown in the rapid scaling plans central to the VCs' roll-up strategies, potentially diminishing the attractiveness of these deals. However, it is widely acknowledged that a few studies are unlikely to significantly deter most Silicon Valley investors. Notably, GC's "creation strategy" companies often begin as profitable ventures, given their acquisition of businesses with existing cash flow – a significant departure from the traditional venture capital model of funding high-growth, cash-burning startups. This profitability is a welcome change for limited partners who have endured years of losses from companies struggling to reach profitability. Bhargava concludes that as AI technology continues to advance and attract substantial investment, an increasing number of industries will become ripe for their incubation efforts.