The Rise of the AI-Powered Service Business

Venture Capitalists have long viewed service businesses with disdain. Marketing agencies, law firms, accounting firms, and other service businesses were fine to employ but would never be attractive investment targets. They require large staffs, have low margins, and are difficult to scale.

VCs prefer to invest in high-margin software businesses. Scaling these businesses requires selling more software — not employing hordes of new staff. Early customers of these software startups would buy into the vision and hope to drive significant ROI by utilizing it.

Generative AI hit the market in 2023. As one would expect, VCs funded businesses that sold software powered by AI to customers who bought into the AI vision. This makes perfect sense, and it is likely that a vast majority of the VC dollars that go into Generative AI businesses will follow this tried-and-true approach.

However, imagine for a second that Generative AI opened up a new path. What if customers were not required to “buy in” to the promise of AI, but rather were able to employ a service business that promised far lower costs and faster delivery than their previous service provider? Examples might include a marketing agency to draft and send marketing emails, an accounting firm to close the books for a startup, or a legal firm to deal with standard contracts. These service businesses could deliver services at a low cost because they themselves have employed AI to drive down human costs in a way that rival “traditional” service businesses could never do.

AI-Powered service businesses may represent an entirely new category of investment for VCs.

Here is why:

1. High Margins: Unlike traditional services businesses, these companies would have high margins because the work traditionally done by humans would be automated by AI.
2. Scalability: Growing these businesses would not require them to grow their staff nor take on the associated overhead (rent, management, benefits) of hiring more people to tackle new demand.
3. Expanded TAM: The TAM for software businesses is typically a fraction of the value of the costs employed by customers to tackle a specific problem. For instance, if someone were to employ software to increase the efficiency of a call center, the value of that software might amount to 10% of spend on customer service (in the best case). However, a call center outsourcing firm powered by AI would be able to charge 5–10x that amount, because it would provide the entire service (not just the software to enhance that service).
4. Simpler Sales Cycles: These service businesses would be a 1-for-1 replacement of an existing service. This would likely result in simpler sales and less evangelizing about the promise of new technology.
5. Roll-up Potential: High-growth software startups don’t typically utilize roll-ups to gain scale. Integrating technologies and customer migrations can create real challenges. However, AI enabled service businesses could utilize the roll-up model to acquire traditional service businesses and dramatically improve margins and operations by implementing AI.

This idea is not entirely new. Prior to generative AI, there were several efforts to automate services. Atrium tried to automate legal services and ScaleFactor tried to automate accounting — both were unsuccessful. However, we believe that the Generative AI boom could make these businesses much more viable, as the underlying technology is vastly superior to what was in the market just a few years ago.

Here are a few examples of categories that could be disrupted by AI-Powered Services.

1. Tax / Accounting / Audit: A mid-sized business might pay an accounting $5k — $50k to close the books every month and at year-end. An AI-powered accountant could process invoices, reconcile/code transactions, and create a draft set of books that would simply need to be QC’d by a small number of accountants employed by this AI-powered accounting firm. We have invested in Finaloop which does this for retail and brand clients. Here is a medium post on why we invested.
2. Legal: Generative AI is quite good at reading and creating legal documents. Many basic legal tasks including contract reviews and basic legal agreement drafting could be completed by an AI-powered legal firm. While this firm would still have a few Partners to oversee the models, it would not have to hire hordes of associates to do the basic review and drafting.
3. Marketing Services: There are thousands of marketing agencies (big and small) that help businesses deploy customers’ marketing dollars. AI could be utilized to optimize a campaign, create segments, and develop creative materials that drive down the cost to serve customers.
4. Debt Collections and Servicing: Financial institutions often outsource debt servicing and debt collection tasks to outside agencies. Many of these tasks and communications could be automated using AI.
5. Countless Others: Virtually any industry that utilizes people-centric agencies to complete repetitive tasks could be a candidate for this AI-powered service model. Further examples include claims processing, call center/customer service, HR/Benefits, underwriting advisory, systems integration, brokerage services, and compliance.