Ben Mathias is a managing partner and Abhijit Gupta is an associate director at Vertex Ventures SEA & India.
Over the last three decades, all kinds of IT Services and BPO businesses have emerged from India, many of which have become multi-billion-dollar companies with worldwide operations.
The reasons for the industry’s success are clear—a large, young, English-speaking workforce, millions of engineers graduating every year, and a sizeable labour cost arbitrage. IT and BPO services currently form a large percentage of India’s exports to the world and these companies have made handsome returns for their investors.
That said, venture capitalists are always on the lookout for promising SaaS companies and IT/BPO businesses hardly feature on their radar.
The reasons are several. The IT/BPO model is often considered more “traditional” than the SaaS (Software-as-a-Service) model. It requires an elaborate recruiting and training engine for fresh college graduates. Enterprise sales cycles are long, following the classic ‘wine-and-dine’ motion.
Custom requirements are the norm, which hinder scalability. In contrast, SaaS selling can be done remotely, often even over a Zoom session.
Once the product is built, it does not require much additional effort except for enhancements and maintenance. SaaS offers predictable revenue streams, with companies achieving 100% year-on-year growth and 80% gross margins, with a less than proportionate growth in team size.
Hence, the scalability and profitability potential have made SaaS attractive to venture capital funding.
The Large Language Model (LLM) has serious potential to bring software-like economics to human-driven services.
Enter AI. The Large Language Model (LLM) has serious potential to bring software-like economics to human-driven services. LLMs can handle unstructured data and contextualise insights, enabling them to be supremely useful in understanding the various nuances of a particular business.
AI agents are evolving from merely assisting with tasks to independently performing them. Automations using LLMs are already replacing (rather than just assisting) many business process workflows and are doing a remarkably good job at it.
Not all tasks however are fully replaceable by AI. While certain functions such as parts of coding and content creation can be handled autonomously, humans remain critical in many other workflows.
For example, in data labelling, humans play an important role in annotating data to train AI models and provide real-world intelligence. This work of the human just cannot be replaced. Even though a system of AI agents can do some parts of this workflow, their cumulative errors often fail to meet the required accuracy.
This makes humans-in-the-loop essential. Highly successful businesses like Mechanical Turk (founded by Amazon) and Scale AI have been built on this concept and are using human labour with automations to train AI models in areas such as autonomous driving.
Today, AI-enabled modules can work in tandem with humans, enabling humans to accomplish work in the same amount of time. An equity research analyst would earlier need weeks to create a research report. Today, she can create one almost in real-time and serve her client better. Similarly, a customer support executive can now handle at least three times the call volume, thanks to voice agents that qualify and route the query to the appropriate human agent.
This is an exciting paradigm shift in the services world, providing a possibility for VC investors to make lucrative returns. At Vertex Ventures SEA & India, we led three investments in line with this theme in 2024. Lenity Health is building an AI + Services platform for US Healthcare.
Attentive AI has built an AI + Services solution for the Construction industry. Our most recent investment blends AI + Services for Security and Guarding. Exporting software and services from India, with AI playing a critical part of enhancing the delivery (improving margins, reducing human effort, improving speed/accuracy), is a theme central to all.
At Vertex Ventures SEA & India, here are some of the key business characteristics we are looking for in early-stage companies building an AI + Services solution:
a. Bottomless market size with a niche wedge: Naturally, there are several massive markets within IT/BPO services. The Big 4 (Accenture, IBM, Deloitte & EY) do $200B in annual revenues, the top 5 Indian IT Services leaders do $90B annually, and McKinsey, BCG, and Bain together do over $35B annually. We are looking for startups which begin by addressing a very specific niche (since AI models perform better on more fine-tuned, richer datasets) and can then expand their value proposition from there.
b. Superior economics and efficiencies compared to BPOs: A human with the assistance of AI, can deliver services faster and with greater accuracy resulting in much higher Gross Margins. While they may not reach the 80% seen in SaaS, they could get as high as 70%. Companies in our portfolio have already achieved these levels. Startups can also quickly scale revenue and achieve profitability through a potentially recurring revenue stream.
They can charge by outcome rather than on the number of humans, which in turn increases their Average Contract Values. For example, Intercom is already pricing their support platform on outcome (per ticket resolution) rather than on input (per human hours).
c. Access to proprietary datasets: Traditional service companies were built on process efficiencies – hiring, training, churning out code and standardizing implementation. AI + Services companies, while delivering the same age-old service can differentiate on better outputs (faster response, greater accuracy, etc.) due to access to proprietary data. AI models can be fine-tuned with datasets that are not publicly available, which makes the service better over time and harder to replicate.
d. Human expertise as a moat to close the loop: Undoubtedly, AI agents will take over more and more workflows that were hitherto done by humans. However, there will still be many areas where humans-in-the-loop are not just a strategic advantage but are also essential. In relationship-driven or trust-based services like wealth management, human involvement fosters the personal connections necessary for client confidence.
Similarly, in regulated industries such as healthcare, compliance often mandates human oversight. In workflows that demand high accuracy such as security operations or financial reconciliation, humans play a vital role in checking AI outputs.
e. Access to the Midmarket: Thanks to AI crushing costs, startups are now able to target Small & Medium Businesses (SMBs) that previously couldn’t afford traditional BPO services. Unlike large enterprise customers, SMBs are used to buying software online. Startups targeting SMBs can enjoy shorter and more repeatable sales cycles compared to those targeting enterprise clients, which would be a more capital efficient and faster growth trajectory for a startup early on before selling to larger clients.
With several key tailwinds as well as inherent moats in the business model, AI + Services is an exciting opportunity for the next set of software and IT entrepreneurs building from India to the world. At Vertex Ventures SEA & India, we are excited to partner with such companies and capture a chunk of this bottomless opportunity.
