AI Software Startups Poised to Dominate $12 Trillion US Services Market

It's no secret anymore that businesses are rapidly integrating artificial intelligence (AI) into their operations. Giants like Salesforce, Hubspot, and Microsoft have already rolled out a variety of AI "agent" features this year, particularly in areas like customer service and sales. But the real excitement is brewing with a new wave of privately funded software companies that are using AI to create groundbreaking applications from scratch. These startups are targeting sectors that have traditionally been slow to adopt technology, such as legal services and healthcare, according to a recent report from Bank of America.

Bank of America's software and services analyst, Alkesh Shah, predicts a boom in AI-native startups over the coming years. In his report dated December 13, based on a virtual conference discussing AI trends, Shah suggests these startups could significantly disrupt the massive $12.3 trillion US Services industry. He compares this emerging scene to the early days of the Internet, saying it feels "like it's 1996 all over again."

Among the startups highlighted at the conference was San Francisco's Hippocratic AI, founded in 2022. They're using large language models to automate non-diagnostic healthcare tasks, such as assessing whether someone needs an emergency room visit. Hippocratic AI's co-founder and CEO, Munjal Shah, shared that their AI agents are not only highly satisfactory but also significantly cheaper, costing between $9-10 per hour compared to $50-90 for human nurses. They claim their AI can outperform human nurses in certain tasks, like identifying unsafe over-the-counter drugs or toxic dosages for specific patients.

Another intriguing startup, vLex from Barcelona, Spain, leverages large language models to generate hypothetical arguments that opposing counsel might use in a lawsuit, helping lawyers and paralegals strategize better. Their software, Vincent AI, also speeds up document searches. Bank of America analysts noted that "the time required to analyze privacy law regulations across seven different countries could be reduced to minutes from weeks." Vincent AI is already popular, with two million users across eight of the world's top ten law firms.

Shah warns that these AI advancements might start encroaching on human jobs. He points out the large numbers of professionals potentially affected, including 3.3 million registered nurses, 55,000 medical scribes, 859,000 lawyers, and 366,000 paralegals and legal assistants. Competing with AI agents, he suggests, could become increasingly challenging.

The Bank of America report also indicates that the use of generative AI is expanding beyond sales and customer service into more industry-specific, "vertical" functions. This shift could be a game-changer for many enterprises struggling to figure out how to best utilize AI technology. The rise of specialized AI software packages might just be the bridge they need to close the gap.