Sell software in the US, build bots in Asia

This article summarizes an episode of Venture Capital’s video series featuring Wally Wang, founding managing partner at Scale Asia Ventures.

Wally Wang, founding partner of Scale Asia Ventures / Photo credit: Scale Asia Ventures

Wally Wang, founding managing partner at Scale Asia Ventures, says most AI advantages are not real. Basic AI models are erasing the benefits of being first. His idea changes how to think about defending a business. The focus shifts from how well an AI works on its own to how deeply it is built into a company’s daily work and whether it has access to proprietary company data.

Wang previously built technology for Microsoft Bing and DataVisor. He later led US tech investments for Fosun, a global conglomerate. He founded Scale Asia Ventures, a venture capital firm focused on enterprise AI. The firm backs startups that connect Silicon Valley’s technical innovation with Asia’s industrial supply chains.

Consumer AI advantage is a myth

Companies are racing to build the next popular AI app for consumers. They assume that being first to market guarantees a lasting advantage. But this plan, which worked for social media and online stores, is not working well for today’s AI.

• No network effects: Unlike older technology, AI apps have not shown strong network effects that lock in users and create a lasting advantage.
• Problems for being first: Building early can create problems with old technology. Companies that come later can use newer, cheaper, and better AI models to copy or improve on the first product’s features.

Wang says the main thing that made past companies successful is missing. “The AI applications are lacking network effects. As you remember, in the Ubers of the world [and] Airbnbs… there are very clear network effects that they have built in. So there’s clearly a first-mover advantage, but in our generative AI applications, we have not seen that moat yet.”

The real advantage is in daily work

If selling AI to people is a weak starting point, companies must look for a lasting advantage by selling to other businesses. Even then, just using a better AI model does not give much of an advantage.

As the basic AI becomes easy for anyone to get, the only lasting advantage comes from becoming an essential part of how a company works every day:

• Focusing on one industry: Companies get an advantage by focusing on one industry. This lets them collect special data that no one else has and build a name that is hard for general companies to copy.
• Becoming part of daily work: The longest-lasting advantage comes from building a product directly into a large company’s daily work. This, plus years of trust, makes it hard for the company to switch to a competitor.

Using the call-center AI company Cresta as an example, Wang explains how being built into a company’s operations creates a real advantage over competitors. “[Cresta] integrates very deeply with the actual call center systems in large enterprises,” Wang explains. “It is not just about the model but also about the integrations into the workflows that operators use day in and day out. The model itself is a commodity, but everything above the model becomes very important, especially for the enterprise.”

The surprising direction of global growth

Once a company has established a defensible workflow, the question becomes one of market expansion. Most observers assume this is a game played best by American companies exporting to the rest of the world.

However, Wang suggests a counter-intuitive reality: it is actually Asian founders who hold a strategic upper hand when entering the US market, capitalizing on a dynamic that American founders struggle to replicate in reverse.

“Everywhere in the world is heavily influenced by the American culture, and people study the US,” Wang observes. “Even if they have not been to the US, they’ve studied the US culturally, economically, and business-wise… The US follows a lot of very transparent and easy-to-work-with rules and policies. The other way around is very hard.”

A plan for robotics companies in different countries