Venture capitalists believe they have found the next major investing edge. Their strategy involves using artificial intelligence to achieve software-like profit margins from traditionally labor-intensive professional services firms. The plan is to acquire mature companies, implement AI to automate tasks, and then use the improved cash flow to acquire even more companies in a roll-up strategy.
Leading this effort is General Catalyst, which has allocated one and a half billion dollars from its latest fund to a “creation” strategy. This approach focuses on incubating new AI-native software companies in specific industries. These new companies then act as acquisition vehicles to buy established firms and their customer bases in the same sectors. General Catalyst has already placed bets across seven industries, from legal services to IT management, with plans to expand to as many as twenty sectors.
The global services market represents sixteen trillion dollars in annual revenue, compared to the one trillion dollar global software market. The appeal of software investing has always been its high margins, as scaling software has little marginal cost. The venture capital thesis is that by automating a significant portion of a services business with AI, the financial math becomes very compelling. For example, in call centers, it may be possible to automate up to seventy percent of core tasks.
The improved cash flow from automation then provides the capital to acquire additional companies at higher prices than traditional buyers can pay, creating a lucrative cycle.
This plan appears to be working. One portfolio company, Titan MSP, received seventy-four million dollars to develop AI tools for managed service providers. It then acquired a well-known IT services firm called RFA. Through pilot programs, Titan demonstrated it could automate thirty-eight percent of typical tasks. The company now plans to use its improved margins to acquire more companies.
Similarly, the firm incubated Eudia, a company focused on in-house legal departments. Eudia has signed Fortune 100 clients like Chevron, Southwest Airlines, and Stripe, offering fixed-fee legal services powered by AI instead of traditional hourly billing. The company recently acquired an alternative legal service provider named Johnson Hanna to expand its reach. The goal for General Catalyst is to at least double the EBITDA margin of the companies it acquires.
General Catalyst is not alone in this thinking. The venture firm Mayfield has set aside one hundred million dollars specifically for investments in “AI teammates.” It led the Series A round for Gruve, an IT consulting startup that acquired a security consulting company and grew its revenue from five million to fifteen million dollars in six months while achieving an eighty percent gross margin.
A managing director at Mayfield stated that if eighty percent of the work is done by AI, a company can achieve eighty to ninety percent gross margins. This could lead to blended margins of sixty to seventy percent and net income of twenty to thirty percent.
Solo investor Elad Gil has also been pursuing a similar strategy for three years, backing companies that acquire mature businesses and transform them with AI. He argues that owning the asset allows for a much more rapid transformation than just selling software as a vendor. Taking a company’s gross margin from ten percent to forty percent represents a huge financial lift.
However, early warning signs suggest this transformation of the services industry may be more complicated than venture capitalists anticipate. A recent study surveyed over one thousand full-time employees and found that forty percent are shouldering more work because of what researchers call “workslop.” This is AI-generated work that appears polished but lacks substance, creating more work and headaches for colleagues.
This trend is taking a toll on organizations. Employees report spending nearly two hours dealing with each instance of workslop, from deciphering it to fixing it. Based on time spent and self-reported salaries, the study authors estimate workslop carries an invisible tax of one hundred eighty-six dollars per month per person. For an organization of ten thousand workers, this could result in over nine million dollars per year in lost productivity. This shows that simply implementing AI does not guarantee improved outcomes.
A leader at General Catalyst disputes the idea that AI is overhyped. He argues that these implementation failures actually validate their approach, proving that it is not easy to apply AI technology to existing businesses. He points to the technical sophistication required, noting the need for applied AI engineers who understand the nuances of different models and how to wrap the technology in software. He believes this complexity is why their strategy of pairing AI specialists with industry experts makes sense.
Still, there is no denying that workslop threatens to undermine the core economics of the strategy. The bigger question is how severe the problem is and whether that picture will change over time.
If companies reduce staff as the AI efficiency thesis suggests they should, they will have fewer people available to catch and correct AI-generated errors. If they maintain current staffing levels to handle the additional work created by problematic AI output, the huge margin gains that venture capitalists are counting on might never be realized.
Either scenario could slow the scaling plans that are central to the venture capital roll-up strategy and potentially undermine the financials that make these deals attractive. However, it will likely take more than a study or two to slow down most Silicon Valley investors.
In fact, because they typically acquire businesses with existing cash flow, General Catalyst says its creation strategy companies are already profitable. As long as AI technology continues to improve, they believe there will be more and more industries for them to incubate companies and apply this model.