The AI services transformation may be harder than VCs think

Venture capitalists believe they’ve uncovered the next significant investment advantage: using AI to bring software-style margins to traditionally labour-heavy service businesses. The emerging strategy centres on acquiring established professional services firms, automating their operations with AI, and then leveraging the improved profitability to acquire additional companies.

At the forefront of this movement is General Catalyst (GC), which has allocated $1.5 billion from its latest fund to a “creation” strategy. This approach focuses on incubating AI-native software companies in specific sectors, which are then used as acquisition vehicles to buy up established firms — and their client bases — within those same industries. So far, GC has made bets across seven fields, including legal services and IT management, with plans to expand to as many as 20 sectors.

“Services globally is a $16 trillion revenue industry,” said Marc Bhargava, who leads GC’s related initiatives, in a recent interview with TechCrunch. “In comparison, software is only $1 trillion globally,” he added, noting that software’s appeal lies in its high margins. “Once software scales, there’s little marginal cost but significant marginal revenue.”

Bhargava argues that if AI can automate 30% to 50% of service company tasks — and even up to 70% in areas like call centres — the financial logic becomes irresistible.

Case Studies: Titan MSP and Eudia

One of GC’s portfolio companies, Titan MSP, has already begun demonstrating this model. With $74 million in GC funding across two rounds, Titan built AI tools for managed service providers and subsequently acquired RFA, a major IT services firm. Through pilot programs, Titan demonstrated its ability to automate 38% of standard MSP workflows, which boosted margins enough to fuel additional acquisitions in a roll-up strategy.

Similarly, GC incubated Eudia, which targets in-house legal departments rather than law firms. The company offers fixed-fee AI-powered legal services, replacing traditional hourly billing. Eudia’s clients include Cargill, Del Monte, and Stripe, and the company has recently expanded by acquiring Johnson Hana, an alternative legal service provider.

Bhargava said GC aims to double the EBITDA margins of the companies it acquires through this model.

The Growing Interest — and Emerging Challenges

GC isn’t the only investor betting on AI-transformed services. Mayfield has carved out a $100 million fund for “AI teammates” startups. One of them is Gruve, an IT consulting firm that bought a $5 million security consultancy and grew it to $15 million in revenue within six months, achieving an 80% gross margin, according to its founders.

“If 80% of the work will be done by AI, you can achieve an 80% to 90% gross margin,” said Navin Chaddha, Mayfield’s managing director, in an interview with TechCrunch. “You could have blended margins of 60% to 70% and deliver 20% to 30% net income.”

Solo investor Elad Gil has been following a similar playbook for several years, investing in firms that acquire mature businesses and then integrate AI to transform operations. “If you own the asset, you can transform it much faster than if you’re just selling software as a vendor,” Gil said in a TechCrunch interview earlier this year.

The “Workslop” Problem

Despite the enthusiasm, early warning signs suggest that transforming services with AI may be more complicated than anticipated. A Stanford Social Media Lab and BetterUp Labs study of 1,150 full-time employees found that 40% are taking on extra work due to what researchers call “workslop” — AI-generated tasks that look polished but lack substance, ultimately creating more work for human colleagues.

Employees reported spending nearly two hours per instance of work — deciphering, correcting, or redoing AI outputs. Based on participants’ salaries and time estimates, the study calculated that workslop imposes an invisible cost of $186 per month per employee. For an organisation of 10,000 workers, this translates to over $9 million in annual lost productivity, according to a Harvard Business Review article.

Bhargava’s Rebuttal

Bhargava rejects the notion that AI is overhyped. He argues that these difficulties actually validate GC’s approach: “It’s not easy to apply AI technology to these businesses,” he said. “If Fortune 100 companies could just hire a consulting firm, sign a contract with OpenAI, and transform overnight, our thesis wouldn’t hold. But in reality, transforming a company with AI is really hard.”

He emphasised that actual transformation demands require AI engineers with experience at companies like Rippling, Ramp, Figma, and Scale, who understand model nuances and know how to integrate them effectively into software applications. This technical complexity, he said, underpins GC’s strategy of pairing AI experts with industry veterans to build new companies from the ground up.

The Margin Dilemma

Still, the issue of workslop could undermine the economics behind these roll-up strategies. If acquired companies downsize to achieve AI efficiency, they risk lacking enough staff to correct AI mistakes. If they maintain current staffing levels, they’ll spend more time fixing flawed AI output, eroding profit margins.

This tension may slow the aggressive scaling plans central to the VC roll-up playbook, although it’s unlikely to deter Silicon Valley investors.

For now, GC’s “creation strategy” companies are already profitable, thanks to acquiring firms with existing cash flow — a significant departure from the usual venture model of funding high-growth, cash-burning startups.

“As long as AI technology keeps improving and investment continues,” Bhargava said, “there will be more and more industries where we can help build and incubate companies.”