a16z Interpretation: What new opportunities is AI creating in the $300 billion BPO market?

Linkloud Introduction

Business Process Outsourcing (BPO) is a $300 billion industry that supports the backend and frontend operations of the world's top 100 large enterprises.

In this episode, a16z podcast host Steph Smith and partner Kimberly Tan take a deep dive into this shift, from call centers and invoice processing to cross-system automation and R&D outsourcing. They discuss how AI is redefining economies of scale, opening up new markets, and extending automation beyond Fortune 500 SMBs. For entrepreneurs and business operators, this dialogue provides a clear idea of the development plan in this new field.

In her article "Deconstructing Business Process Outsourcing: How AI Will Disrupt Outsourcing Work," Kimberly Tan explores in depth and comprehensively how the rise of AI challenges the status quo. Article link:

Key Takeaways:

The human-centered BPO model has inherent limitations; one person cannot handle 100 tasks simultaneously, and humans may produce misunderstandings and errors due to various external factors.

The best opportunities for startups are those scenarios with extremely clear ROI, typically choosing those types of features that have clear KPIs to evaluate work effectiveness.

The impact of AI on the BPO industry is not just about replacing existing manpower; it is doing something even more important: expanding the boundaries of the entire market.

The development of AI Coding is not about directly seizing existing BPO business, but rather empowering end users to enable them to solve problems on their own.

AI is quietly reshaping a $300 billion invisible empire.

As the wave of AI sweeps across the globe, while we discuss LLM, AIGC, and those cool consumer applications every day, there is a vast and "invisible" industry that is being fundamentally reshaped by the power of AI. This industry is business process outsourcing, commonly referred to as BPO.

BPO: Ubiquitous Service

BPO services have actually penetrated all aspects of our daily lives. When we call bank customer service, communicate with airlines about tickets, or receive after-sales support from e-commerce, the person serving us on the other end of the phone is very likely from a BPO company.

Accenture, Tata, Wipro, Cognizant, Infosys, these globally renowned IT and consulting giants have a significant part of their business landscape composed of BPO.

The core essence of BPO: Reducing the burden for enterprises

In simple terms, BPO refers to when a company reaches a certain scale, there will be a large number of non-core but essential business processes. If all of these are managed by internal teams, it becomes very complex and costly.

Therefore, companies choose to outsource these tasks to specialized institutions in order to achieve higher efficiency and scalability.

These tasks not only include the customer support and customer service we are familiar with, but also cover a large number of unseen backend functions, such as:

Outsourced HR

Financial Accounting

Invoice processing

Knowledge management and research work

The role of BPO: the "lubricant" of the modern business world

This industry is like the essential gears and lubricants in the vast machinery of the modern business world, ensuring the smooth operation of the entire system. Its history is quite long, dating back to the 1940s when some companies began to help manufacturing enterprises manage their complex operating processes.

As of today, it has deeply integrated into almost all major industries involved with the Fortune 500 companies—retail, tourism, telecommunications, logistics, manufacturing, healthcare, insurance, and banking.

This is a massive industrial ecosystem, with a current market value of 300 billion dollars, and it is expected to grow to an astonishing 500 billion dollars or more by 2030.

Its continuous growth itself illustrates the enormous amount of work that large enterprises need to accomplish to maintain their daily operations. However, this gigantic empire built on the foundation of "human labor" is now facing an unprecedented disruption. The force of this disruption is AI.

The essence and limitations of the traditional BPO model: Starting from "people"

To understand how profound the disruption brought by AI is, we first need to see clearly the essence of the traditional BPO model and its inherent limitations.

Although BPO companies also provide services such as strategic consulting or outsourced application development, the focus of our discussion is on the most core business process outsourcing tasks that are executed by people.

Inherent problems of traditional models

The core of the traditional BPO model ultimately comes down to "people." Whether it is customer service representatives answering calls or accountants processing invoices, it is specific individuals who ultimately complete these tasks.

This means that the efficiency, quality, and scalability of this multi-billion dollar industry are constrained by the limitations of humanity itself.

One of the most obvious problems is the delay. A person cannot handle 100 things at the same time. During peak periods of customer support, users often have to wait for a long time to connect to a human agent, an experience that many people can relate to.

Another issue is the deviation and errors in understanding. When humans handle repetitive and procedural tasks, they inevitably make mistakes due to fatigue, emotions, or lack of experience.

Although companies outsource these tasks because they are not their core competencies or they do not want to expend effort managing them, this does not mean that these tasks are being performed in the best possible manner.

In fact, many enterprises are well aware of this, but for a long time, they have not had better options.

Limitations of traditional software: Unable to replace human cognitive abilities

Why was software unable to solve this problem in the past? Because the historical limitations of traditional software lie in its proficiency at executing processes that are clearly defined and relatively stable. It requires structured data input, clear rules for instructions, and does not need to understand and judge within complex contexts.

However, most of the work handled by BPO is precisely the "forbidden zone" of traditional software.

In a customer service scenario, you must truly understand what the customer is asking over the phone, whether their tone is anxious or confused. For example, when handling an invoice, you must be able to understand the content of the different sections on the invoice, even if their formats vary widely.

This type of work is filled with unstructured information and scenarios that require instant judgment, which traditional software is simply incapable of handling. Therefore, the only solution is to invest a large amount of human labor to compensate for the shortcomings of software with human cognitive and judgment capabilities.

This is the fundamental reason for the existence and development of the traditional BPO industry. It uses a "human wave tactic" to solve a problem related to unstructured information processing and cognitive judgment that software cannot resolve. However, this solution also constitutes its greatest ceiling.

A New Paradigm Emerges: How AI Breaks the Chains of Software

The emergence of AI does not merely bring about a simple optimization of traditional software, but rather a fundamental paradigm shift. AI, especially LLMs, is truly adept at handling the issues that have historically troubled traditional software.

It is very good at acquiring massive amounts of usually unstructured, variably formatted information dispersed across different systems, then synthesizing, constructing, and understanding that information, and ultimately, based on that understanding, outputting a targeted action.

Voice AI: A Breakthrough from Zero to One

In this AI-led transformation, some technologies have demonstrated incredible ROI, becoming pioneers in disrupting the BPO industry, while others herald a broader future.

Among them, the capabilities of Voice AI have achieved a decisive breakthrough from zero to one. We have all had the unpleasant experience of dealing with traditional telephone robots, getting lost in complex voice menus, or being repeatedly tormented by a rigid voice that completely fails to understand our intentions.

However, the technological innovations of Voice AI in recent years have been astonishing. Nowadays, you might be able to have a natural conversation with an AI agent on the other end of the phone, so much so that at the beginning of the conversation, it's hard to distinguish whether the other party is a real person or AI.

They not only sound very human in their conversational abilities and tone, but the latency has also become very low, allowing them to respond at a normal human response speed.

More importantly, these AI agents can connect to the business systems of enterprises, allowing them to understand your background information during calls and provide quicker and more accurate responses.

Future breakthroughs: The potential of "operator" technology

If Voice AI is an application that has already been implemented, then another emerging technology suggests that AI will unlock a broader world of automation. This technology can be referred to as "Operator" or browser usage technology.

The core idea is to enable AI agents to work like humans across various software systems and interfaces, whether traditional desktop software, web applications, or custom systems within enterprises.

Soon, AI agents will be able to navigate these complex applications, not only to gather information but also to take action. This means that jobs that used to require human intervention, such as data analysts or invoice processors, can be handled by AI in the future.

New vs. Old Showdown in the Arena: Where Are the Opportunities for Startups?

In the face of such a tremendous transformation, traditional BPO giants like Accenture and Tata certainly will not choose to sit idly by. They have a deep understanding of the opportunities brought by AI. For startups, there are also genuinely exciting windows of opportunity in the short term.

This mainly stems from two core reasons:

First, there is a fundamental difference in the business models. The business model of these traditional BPO giants is based on "labor force," relying on a large number of personnel to execute tasks.

For any large publicly traded company with annual revenues in the hundreds of billions of dollars, shifting the core business from "people" to "AI products" is an extremely difficult and painful transition. This transformation is bound to be slow.

Secondly, many people underestimate the practical difficulties of working with these advanced AI systems. You need to invest a significant amount of effort to ensure that the AI does not produce "hallucinations," establish a set of evaluation criteria to assess the quality of the AI agent's responses, and also determine when and how to replace the underlying model.

You must be a truly AI-native technical founder to deeply understand how to navigate these complexities, and this is far from a widely distributed skill today.

High ROI scenarios: Excellent opportunities for customer support

Therefore, the best opportunities for startups are those scenarios where the ROI is extremely clear, typically involving features that have specific KPIs to assess work effectiveness.

Customer support is a perfect example. Its KPIs are very clear: how many tickets can be handled within a certain time frame, and the satisfaction score (CSS) of the end users after the tasks are completed.

You can clearly demonstrate the value of AI agents with data. In contrast, KPIs in some areas, such as HR, are much more vague. Of course, this also requires more work to persuade companies to quantify the value of an AI HR assistant.

In addition, AI is not omnipotent and will always have some very long-tail, extremely complex or special problems that require human wisdom to address. This means that successful business models must consider who will handle these long-tail tasks that AI cannot solve.

Expansion and Creation: AI is not only reshaping but also pioneering new markets.

The impact of AI on the BPO industry is not just about replacing existing manpower; it is also doing something more important: expanding the boundaries of the entire market.

In the past, BPO services were mainly prepared for large enterprises with sufficient budgets. The emergence of AI solutions has enabled small and medium-sized enterprises that had never been in contact with BPO to also enjoy such services.

For example, a small to medium-sized e-commerce company may have previously been unable to afford the cost of a 24-hour customer service team, but now by deploying an efficient AI customer service agent, they can actually provide round-the-clock support to their customers at a very low cost.

This model may not directly impact the core business of those large BPO companies in the short term, as these small and medium-sized enterprises are not their target customers. It is equivalent to opening up a brand new incremental market.

But in the long run, if these startups providing AI solutions grow alongside their customers and gradually target larger clients, they will ultimately have a profound impact on the existing market landscape.

The value enhancement of AI for existing BPO clients

For companies that are already using BPO services, AI is also helping them expand the range of services that can be covered.

In the past, a company might have only provided human support on its core product line, but now, with the help of AI, they can extend high-quality customer service to every corner of their entire product line.

A good way to identify these best opportunities is to look for operational roles that "scale linearly" with the growth of the company. This means that the costs of this work will increase proportionally with the growth of the company's business.

The more customers you acquire, the more customer support requests you will have; the larger your business grows, the more invoices you need to handle. If you can provide an AI solution that flattens this steep cost growth curve or even allows it to decrease, this would be an extremely clear and irresistible value proposition for the company.

When AI makes everyone a developer

Many large BPO companies do a lot of work that not only involves outsourcing business processes but also includes outsourcing IT or application development. They may also build some small internal tools or applications for companies that do not have internal IT or engineering resources.

Although building a complete application is much more complex than responding to a customer inquiry, a larger trend is that AI Coding Agents are becoming increasingly better.

The development of this technology will empower those who may not be as technical, or even completely lacking a technical background, to build the complete applications they need.

This will actually become a very interesting "orthogonal attack vector" targeting the BPO industry. The way it disrupts is not by directly seizing the existing business of BPO companies, but by empowering end users, enabling them to solve problems on their own, thereby making the need for "outsourcing" itself disappear in certain scenarios.

It is difficult for us to quantify what this really means in the next two to three years. But one can imagine the kind of transformation that will occur when you truly enable a brand new demographic to build their own applications. This is not only a reshaping of an industry, but also a liberation of the way we work and creativity itself.