Conversion of AI services may be more difficult than VCS thinks

Venture capitalists are sure they have found their next big investment advantage. This uses AI to narrow down the margins like software from traditional labor-intensive services businesses. This strategy involves acquiring mature professional services companies, automating tasks to implement AI, and rolling up more companies with improved cash flow.

The general catalyst (GC) leads the bill. This is dedicated to the latest fundraising called the “creation” strategy, which focuses on incubuted AI native software companies in a particular industry, and to those that focus on purchasing established companies and their customers in the same sector using the company as an acquisition vehicle. GC has bets on seven industries, ranging from legal services to IT management, and plans to fully expand to up to 20 sectors.

“Services worldwide are revenues of $16 trillion a year across the globe,” said Marc Bhargava in a recent interview with TechCrunch, which leads GC’s related efforts. “By comparison, software is only $1 trillion worldwide,” he said, adding that the appeal of software investments is always a higher margin. “As you scale your software, there’s very little marginal costs and a lot of marginal revenue.” If you can automate your service business, he said – working on 30% to 50% of companies with AI, and in the case of call centers, even automating up to 70% of those core tasks, maths will start to look attractive.

Improved cash flow provides ammunition to acquire additional companies at a higher price than traditional buyers can afford, creating what their supporters consider to be a lucrative flywheel.

The game plan appears to be working. Check out Titan MSP, one of General Catalyst’s portfolio companies. The investment company provided $74 million over two tranches to help develop AI tools for managed service providers, and later acquired RFA, a well-known IT services company. According to Balgava, through the pilot program, Titan has demonstrated that 38% of typical MSP tasks can be automated. The company is currently planning to use improved margins to obtain additional MSPs with classic rollup strategies.

Similarly, the company incubated Eudia, which focuses on the internal legal department rather than the law firm. Eudia has signed up for Fortune 100 clients, including Chevron, Southwest Airlines and Stripe, offering fixed-rate legal services powered by AI rather than traditional hourly billing. The company recently acquired Johnson Hanna, an alternative legal service provider, to expand its reach.

General Catalyst expects it to be at least twice as much.

This is not the only way of thinking in a major country company. Venture company Mayfield led Series A for IT consulting startup Gruve, which has specifically opened up $100 million for its “AI Teammates” investments, acquired a $5 million security consulting firm and achieved a 80% total margin while increasing revenue to $15 million within six months.

“If 80% of the work is done by AI, you can get 80% to 90% gross profit,” Mayfield’s Managing Director Navin Chaddha told TechCrunch this summer. “We were able to blend a margin of 60% to 70% and generate a net profit of 20% to 30%.”

Solo Investor Elad Gil has been pursuing a similar strategy for three years, helping companies acquire mature businesses and transform their AI. “If you own an asset, you can [transform it] In an interview with TechCrunch this spring, Gill is much faster than if he were selling software as a vendor. “And that’s a big lift because you’re getting a total company margin, for example, 10% to 40%.”

However, early warning signs suggest that this whole service industry pervert may be more complicated than VCs expect. A recent survey by researchers at Stanford Social Media Lab and Betterup Labs found that 1,150 full-time employees have been surveyed across the industry. 40% of these employees are doing more work, refined, but without substance, producing more work (and headaches) because it is what researchers call “Workslop.”

This trend is hitting the organization. Employees involved in the investigation say they spend an average of nearly two hours dealing with each instance of Workslop, deciphering it first and deciding whether to send it back and whether to send it back just to fix it.

The study authors estimate that worklops have invisible taxes of $186 per person per month, based on estimates of the time spent along with their self-reported salary. “Given the estimated prevalence of workslop, for an organization of 10,000 workers, this is a decline in productivity, exceeding $9 million a year,” they wrote in a new Harvard Business Review article.

Implementing AI simply does not guarantee improved results.

Bhargava challenges the notion that AI is exaggerated and instead argues that all of these implementation obstacles actually test the general catalytic approach. “I think it shows an opportunity that applying AI technology to these businesses is not easy,” he said. “If all Fortune 100 and all these people can just bring in consulting companies, slap some AI, sign a deal with Openai and transform their business, then obviously our paper [would be] It’s a little less robust. However, the reality is that changing a company with AI is really difficult. ”

He pointed out that the technical refinement required in AI is necessary as the most important missing puzzle piece. “There are a variety of technologies, and it’s great at a variety,” he said. “We really need these applied AI engineers from places like rip rings and ramps, figmas and scales, who work with different models to understand the nuances, what suits what, and how to wrap them in software.” The complexity, he argued, is why Catalyst’s general strategy of combining AI professionals with industry experts makes sense.

Still, there is no denying that Workslop is threatening to undermine the core economics of strategy. The bigger question is how serious the problem is, and whether the picture changes over time.

For now, when companies reduce staff, as suggested by AI efficiency papers, fewer people will be able to catch and correct errors generated by AI. If you maintain current staffing levels to handle additional work created by the problematic AI output, the large margins your VCs rely on may not be realized.

It is easy to argue that both scenarios may slow down the scaling plan, which is central to VCS’ rollup strategies, and potentially undermine the numbers that make these transactions attractive. But let’s face it. Slowing down most Silicon Valley investors requires one or more research.

In fact, they usually buy companies with existing cash flow, so General Catalyst says that its “creation strategy” company is already profitable.

“As AI technology continues to improve and as long as we see this massive investment and improvements to our models, we think there will be more and more industries that will help businesses stay in hiding,” Bhargava said.