Everyone in SF knows we can build AI agents to productize a service—but what next? How do you turn that into a real business?

I grew up in a family of tech consultants and service providers, witnessing everything from medical billing to software implementation. There are good reasons why these ecosystems of service providers exist.

In this blog, I’ll share why some of these systems are built the way they are, the nuances of productizing them, and what we’ve learned while building a company in this space.

Service Businesses Look Like Onions

As an engineer in SF, if you closely examine a services business—peeling it layer by layer—you might wonder why it even exists. In IT services, a U.S. company hires a U.S. firm, which works with an Indian partner, which in turn works with a staffing agency. The engineer ends up with four email addresses, one for each layer.

In insurance, a company hires an advisor, who finds a TPA, which has a back office run by yet another firm. Most of the money in this system is made by middlemen who specialize in finding the "right" service provider.

In my view, finding the right service provider is an NP-hard problem—the service industry has been solving it through brute force.

Will the Traditional SaaS Playbook Work?

The SaaS playbook is clear: find a niche problem, build a product, and expand around it. If the starting point is right, you can build a $10B company. In services, there are countless pain points that could become software products. But applying YC-style thinking often leads to solving problems that won’t exist in a year.

For example, IT services firms spend time writing proposals and requirement docs—tasks that AI can easily automate. But do they need to be automated? If you build an AI-native Salesforce implementation firm today, you wouldn’t follow the traditional approach of documentation and user stories. You’d focus on solving core business problems like broken sales processes or inefficient quoting.

Go AI native from Day 1

Is the Sky the Limit?

A service company takes on whatever comes its way. But in a productized service, where do you draw the line? This is a tough call—and where the founder’s gut instinct matters most. Go too narrow, and you become a point solution. Go too broad, and no one understands what you do.

A friend of mine is building an AI agent for policy consulting, focused on tech service firms like TCS and Wipro. When a large bank onboards a vendor, the vendor must create a policy document outlining rules for contractors (e.g., no phones at work). Today, this is done manually by policy consultants using industry regulations and company policies. Automating this with AI is a perfect niche—clear value, repeatable workflow, and a path to expand into other types of policy document creation.

Fancy AI Rollups

Buying a firm is like getting married—you need to go on a few dates first. It’s tempting to say we’ll acquire services firms, replace their tech with our AI agents, and boost margins. But in reality, this takes careful planning and execution.

As a first step, we're working with PE firms that focus on IT services—firms we're aiming to replace—by joining their incubators and startup programs. This gives us a path to build relationships with target service providers by first selling our tech. Once trust is established, we can leverage the PE firm’s expertise to pursue a proper acquisition.

It’s a two-year roadmap, and we’re just getting started. There’s a lot of blood and sweat ahead to make it work.

Do Not Innovate

We believe there should be no innovation in pricing—because customers hate it. Instead of presenting 100 variables for a CRO to forecast, we use simple heuristics and provide a quote based on a 2–3 week pilot. It works like magic.

Ending Note

This is an exciting time to be in this space. If you’d like to connect with industry veterans (some of whom are my extended family 🙂) or startup founders working on similar problems, feel free to shoot me an email at abhishek@ressl.ai.