Venture capitalists are increasingly convinced they’ve discovered a new investment advantage: leveraging AI to extract software-level profit margins from sectors that have traditionally relied heavily on human labor. Their approach centers on acquiring established professional services companies, automating processes with AI, and then using the resulting financial gains to purchase additional firms.
General Catalyst (GC) is at the forefront of this movement, allocating $1.5 billion from its latest fundraising round to a “creation” strategy. This approach focuses on building AI-first software businesses within targeted industries, which then serve as platforms to acquire other established companies—and their client bases—in those same fields. GC has already invested in seven sectors, including legal and IT services, and aims to eventually reach as many as 20 industries.
“Globally, services generate $16 trillion in annual revenue,” said Marc Bhargava, who leads GC’s initiatives in this area, during a recent conversation with TechCrunch. “By comparison, the global software market is just $1 trillion,” he pointed out, emphasizing that software’s appeal has always been its superior margins. “Once software scales, the incremental cost is minimal, while additional revenue can be substantial.”
He explained that if services companies can also be automated—targeting 30% to 50% of their operations with AI, and up to 70% for core tasks in areas like call centers—the financial benefits become highly compelling.
This strategy appears to be yielding results. For example, Titan MSP, a company in General Catalyst’s portfolio, received $74 million in two funding rounds to develop AI solutions for managed service providers. Titan then acquired RFA, a prominent IT services provider. According to Bhargava, pilot projects showed Titan could automate 38% of standard MSP tasks. The company now intends to leverage its improved profitability to acquire more MSPs, following a traditional roll-up model.
In a similar vein, GC helped launch Eudia, which targets corporate legal departments instead of law firms. Eudia counts Fortune 100 companies like Chevron, Southwest Airlines, and Stripe among its clients, offering AI-driven legal services for a fixed fee rather than hourly rates. The company recently expanded by acquiring Johnson Hanna, an alternative legal services provider.
Bhargava noted that General Catalyst aims to at least double the EBITDA margins of the companies it acquires.
Other firms are adopting comparable approaches. Mayfield, for instance, has set aside $100 million for investments in “AI teammates,” such as Gruve, an IT consulting startup. Gruve acquired a $5 million security consulting business and grew its revenue to $15 million in just six months, achieving an 80% gross margin, according to its founders.
“If AI can handle 80% of the workload, gross margins could reach 80% to 90%,” said Navin Chaddha, Mayfield’s managing director, in a summer interview with TechCrunch. “This could result in blended margins of 60% to 70% and net income between 20% and 30%.”
Solo investor Elad Gil has been following a similar path for three years, supporting companies that acquire established businesses and overhaul them with AI. “Owning the asset allows for much faster transformation than simply selling software as a vendor,” Gil told TechCrunch in a spring interview.
However, early indications suggest that transforming the services sector may be more challenging than VCs expect. A recent study by Stanford Social Media Lab and BetterUp Labs, which surveyed 1,150 full-time employees across various industries, found that 40% of respondents are burdened with extra work due to “workslop”—AI-generated output that looks polished but lacks substance, creating more tasks (and frustration) for colleagues.
This trend is affecting organizations negatively. Surveyed employees reported spending nearly two hours on each instance of workslop—first to interpret it, then to decide whether to return it, and often just to correct it themselves.
Based on participants’ time estimates and self-reported salaries, the study’s authors calculated that workslop imposes a hidden cost of $186 per employee each month. “For a company with 10,000 staff, given the frequency of workslop, this translates to more than $9 million in lost productivity annually,” they wrote in a recent Harvard Business Review article.
Bhargava pushed back against claims that AI is overvalued, contending that these implementation challenges actually support General Catalyst’s methodology. “I think it highlights the opportunity—it’s not simple to apply AI to these businesses,” he said. “If every Fortune 100 company could just hire a consulting firm, add some AI, sign a deal with OpenAI, and revolutionize their operations, our thesis would be less compelling. But in reality, transforming a business with AI is extremely difficult.”
He emphasized that the technical expertise required for AI is the key missing element. “There are many different technologies, each with its own strengths,” he said. “You need applied AI engineers from companies like Rippling, Ramp, Figma, and Scale—people who know the models, their nuances, and how to integrate them into software.” This complexity, he argued, is precisely why General Catalyst’s model of combining AI experts with industry veterans to build companies from scratch is effective.
Nonetheless, it’s clear that workslop could undermine the financial logic behind this strategy. Even if a holding company is established, if acquired firms reduce headcount as AI efficiency suggests, there may not be enough staff to catch and fix AI errors. If they keep staffing levels steady to handle the extra work from flawed AI output, the anticipated margin improvements may never materialize.
These issues could, in theory, slow down the aggressive expansion plans that are central to the VCs’ roll-up strategies and cast doubt on the financial projections that make these deals appealing. But realistically, it will take more than a few studies to deter most Silicon Valley investors.
In fact, because General Catalyst typically acquires businesses that already generate cash flow, its “creation strategy” companies are profitable from the outset—a significant shift from the usual VC model of funding fast-growing, loss-making startups. This is likely a welcome change for the limited partners who have long supported venture firms through years of unprofitable investments.
“As AI technology keeps advancing and we continue to see significant investment and progress in these models,” Bhargava said, “I believe there will be even more industries where we can help launch new companies.”
