A while back, a CEO introduced me to his CTO. Big company, well over a billion in revenue. I walked him through the whole Qurrent story, and when I finished, he said, “I’m not your buyer.”

It wasn’t a brush-off. He understood exactly what we do. “You should be talking to our operators,” he said. “I’ll weigh in, but this isn’t a technology decision. It’s an operating decision.”

That conversation has repeated itself more times than I can count. The people who get what we do, and love it, are the CFOs, VPs of finance, and VPs of operations. CIOs and CTOs tend to either want to build something in-house or feel it’s simply not their call to make. It took us a while to find the right words for why.

The market only has two boxes

We went live with our first enterprise customers in early 2024, custom building agents and then managing them on their behalf across marketing, customer support, operations, and finance. We were one of the very first companies doing this. And we had a hard time explaining it, because the market really only understands two kinds of vendors. You’re either software or you’re services. We’re something different.

Software was the box most people reached for. They’d send us to IT as if we were selling a tool. But a SaaS provider hands you software, and it’s up to you to make it work and own the results. We take the whole function off your hands. If something goes down in the middle of the night, or the performance isn’t quite right, it’s on us to jump in and fix it immediately. That’s also why you hear so much about outcome-based pricing in this space right now. Qurrent isn’t alone. Many other AI agent companies are also selling outcomes, not software. We’re not selling software. We’re selling outcomes.

Services didn’t fit either. Services usually means consultants and headcount. We’re not renting out people. We’re running the function with agents and we’re accountable for the result. We tried calling it a managed service for a while, but you could argue SaaS is a managed service too, just managed by software. None of the existing labels captured what was really happening: companies were outsourcing entire operations to us.

The closest thing the market already understands is BPO. The analogy isn’t exact. A traditional BPO firm manages a large team of workers in India or the Philippines. We manage agents. But the relationship is the same. The enterprise hands over a function, and the provider is accountable for it.

A $330 billion blind spot

Here’s what surprised me once we adopted the frame. BPO was roughly a $330 billion industry globally in 2025, nearly as big as the entire SaaS market. Yet almost nobody in tech talks about it. For years, services was a bad word in venture capital. Investors wanted the supposedly infinite scalability of software, so when agents arrived and everyone recognized how powerful they were, a lot of companies made the natural mistake of trying to productize them.

That doesn’t work. Agents are workers. They have to be tuned to a company’s backend systems, its reasoning, and its specific business processes. You can build a reusable framework and productize certain components, but that last 20% has to be customized, and the forward deployed engineer is essential. Plenty of companies resist that because they’re worried about scalability. The irony is that this model scales far better than traditional consulting or traditional BPO.

So here’s how I’d define it: an agentic BPO is a service provider that takes on entire business functions for enterprises, performs that work with agents, and owns the outcome.

Why now, and why finance

Over roughly the last year, the underlying models crossed a threshold in reasoning. Some simpler use cases were ready earlier, and some still aren’t there. But on the whole, we can now take on genuinely complex business processes, even in finance operations, where the numbers have to be right and it’s real dollars on the line.

We didn’t plan to go into finance. With one of our largest customers, a global digital media company, we started with an advertising use case. With another, one of the country’s largest insurance companies, it was sales. We went from signed contract to live production quickly, and in both cases the next conversation was the same: we’ve got a real pain point, can you tackle this? Both times it was finance operations. That’s the best signal you can get, when your customers are pulling you in a direction.

Finance is the hardest use case there is. The workflows are complex and there’s no room for error. Because we’ve been doing this longer than anyone and our technology is the most mature, we’re able to succeed where others can’t. That’s why finance operations is now our focus.

Bigger than BPO

Finance and accounting is the second-largest slice of BPO, at $70B per year, yet many companies still keep it in-house. Traditional outsourcing doesn’t always provide a great customer experience. It’s error-prone, visibility is limited, and it’s so people-intensive that changing a process can mean months of change management.

Agentic BPO flips all of that. Change the instructions and every agent behaves the new way immediately, so the business becomes far more nimble. Error rates drop. And you can see every decision every agent ever made, and exactly why, in a form a business user can audit without calling an engineer.

When Uber started, it wanted a share of the taxi market. Instead, it grew far bigger than the taxi industry ever was, because far more people wanted rides once the experience got better. Agentic BPO will do the same thing to BPO.

That CTO I mentioned saw it before most of the market did. This is an operating decision, and the companies that treat it that way are going to set the pace.