The venture capital world has been built on exceptions rather than rules.
Less than 1% of all new businesses secure venture capital.
The rule is: most businesses don't get venture capital, period.
Taking it a step further, around 70% to 80% of venture capital investments usually lose money.
So… making money in venture capital is about finding the 20% - 30% of exceptions, within the less than 1% of exceptions. It’s a case of exceptional investors putting money into exceptional founders with exceptional businesses in the hopes of exceptional returns. Ironically, to be any good at finding these exceptions, venture capitalists need high quality rules. They need to know what the highest-performing exceptions have in common.
This is one reason AI is inherently disruptive in the world of venture capital. VCs can be great at identifying patterns and creating mental rules of thumb, but pattern recognition – especially when it comes to large data sets and mathematical complexity – is where AI takes things to another level.
A shift away from pure human subjectivity and bias, towards more objective rules and patterns, has some obvious benefits in venture capital. In addition to better pinpointing which startups to invest in, advanced pattern recognition can also facilitate more efficient VC operations by helping to streamline just about every step of the investment process while enabling more productive and scalable investment strategies.
However, more reliance on predefined rules poses challenges for startups that deviate from the norm.
Exceptional opportunities that fall outside the established AI criteria may be overlooked, hindering innovation and potential breakthroughs. In this way, AI's inability to recognize unconventional but promising ventures might limit or homogenize the startup ecosystem.
Keep this in mind, but also keep in mind...
Good AI can identify rules, but great AI can identify exceptions too.
In the world of venture capital, an ability to identify great startups that other investors overlook can be a huge competitive advantage. This creates an incentive for VCs to use AI that not only finds static patterns and rules, but to also embrace systems that are dynamic and capable of evolving under real-world circumstances. This gives at least some hope that more AI in venture capital won't necessarily lead to an overly-constrained, homogenous innovation landscape. In fact, AI may even create more diversity in the startup ecosystem. In practice, we’ve certainly found this to be true – every day our AI points us to innovative startups we never would have otherwise been aware of or considered.
Today, the state of AI in venture capital is a mosaic of diverse AI and human-hybrid flavors, creating optimism that, even with the rise of AI, there will continue to be diversity in startups, not to mention diverse metrics, models, and decision-making strategies.
Perhaps the most disheartening impact of AI in venture capital (at least to some founders) is the degree to which it strips away a fundraiser’s ability to bullshit.
AI and data can make it harder for founders to rely on traditional salesmanship, relationships or emotional appeals. While a bummer for some founders, this shift levels the playing field by emphasizing objective business performance and quantifiable metrics. Too many otherwise undeserving founders have raised money using insider hoodwinks, charisma or emotional pitches. Similarly, AI can take steps to mitigate the effects of socio-economic, ethnic, gender or other forms of discrimination and bias in ways that simply aren’t fathomable when decisions are instead made by groups of humans based on largely inarticulate, collective gut feelings.
The integration of AI into venture capital represents a world shaped by rules that look for exceptions.
The shift towards objective rules and away from human subjectivity is good for venture capital to the extent that it offers greater accuracy, efficiency, scalability and the potential for a more inclusive ecosystem. Yet, the caveat remains that startups falling outside AI's criteria might be overlooked. A clever paradox lies in AI's ability to excel at identifying the best exceptions, even to its own rules.