Human beings are always proud of what they create. They like creating things and then teaching them how to exactly be like them. So much so for leaving a legacy.
Charles Babbage, the widely known creator of the first mechanical computer, wouldn’t have, in his wildest dreams, thought about something known as artificial intelligence. And neither would have Henry Kravis, about investing in this space, 248 years later. Artificial Intelligence is poised to become a transformative force in human history, and we are about to witness a whirlwind change in the way we interact with machines. The once feared, and potentially problematic disadvantage of repeated bean counting computing devices is now being made to climb a Lorenz curve, and suddenly, the times ahead are exciting.
Simply put, artificial intelligence is machine learning. That is training the machines to think and make decisions just as the way we do in our daily lives, with a one tiny twist – all the decisions are empirical or based on analyzing the base data. Over the last 4 years, big data and data analysis along with data scientifics have caught our fancy. Tech has seen a lot of activity in the arena and startups offering big data solutions have sprung up more than the ones delivering online groceries. And as a result, we have changed from individuals to simple data points, with trackers on every single activity of ours. If we walk a step, the machine knows; if we eat at a restaurant, the machine knows; the machine knows the identities of our families, friends, where we go, what we do, what we look at, and what we buy. And based on each one of those, we have big data companies tracking our moves and roping in digital advertising to make us do all that we do, online. This revolution where there are concerns raised over digital and actual privacy, and disabling the location services on our phone has become a fad, we lead two lives – offline and online.
And it’s about time that our online version gets at least as smart as our offline persona.
Investing in artificial intelligence is quite a challenge. It adds a layer of complexity to the already complex valuations of futuristic companies. That being said, there are quite some investments happening around the world in this area. We have all heard of Google and Facebook acquiring a string of companies which show even the tiniest of potential, and customizing the content that shows up in our search results or news feed respectively; but apart from the obvious two, there are several companies buying AI driven algorithms to power their online offering. If such is the kind of activity happening in this space, the traditional investors aren’t too far behind. The investment from PE/VC in AI companies looks nothing short of the ticket sales graph of the new Star Wars flick.
AI has astounding 293% in YoY funding growth in 2014. This would further explain the magnitude of change.
Anyway. Keeping it short, the point of this article is how with the use of AI, the alternative assets industry can change the way it looks at investments. Let’s dive straight into it.
PE/VC – The way it is.
The alternative asset management industry is huge. And practically controls most of the companies on this planet. Everything in this industry revolves around deals and the way they are structured. Now these deals are finite window transactions, which implies there is a premium to every bit of delay that so happens. Relying entirely on excel sheets, which aren’t tamper proof, the decisions are based on a lot of financial models that give the targeted return on investment, at least on paper. While there isn’t a clear pattern in the consequent decisions, but there is a loose pattern, commonly referred to as a fund strategy. The decisions on where and how to invest remain restricted to the fund manager’s mind. The diligences take up most of the time in the deal cycle, slowing the process down. The process has such gaps that an entire bridge loan industry and funds sprung up to cover for the companies’ need for capital! While most of the decisions are spot on, the point here is that a lot of time gets wasted in checking the boxes and crossing the lines on a whiteboard. And time is precious.
That’s the story of how it exists today.
Why is it that the people who fund the technologies and drive the business revolutions in the world, remain aloof from the technology themselves?
What can we do?
AI can transform the way this whole process works. By infusing technology with on ground knowledge, and certain deep learning mechanics, the deal team can be broken down to a single person capable of running queries and enabling the entire transaction seamlessly. The idea here is to critically match the evaluation of a deal with an intelligent software capable of handling the nuances. Not only would the time required to close go down, but also the error rate, and we will be closer to the expectation of making fully informed decisions.
One of the primary lessons in microeconomics on the type of markets is on a perfectly competitive market, one in which all the stakeholders have equal amount of information. Can we ever reach such a state with AI in PE? Can machine learning actually change the industry? How long would it take for this revolution to be materialized? Introduction of this technology definitely opens doors for steadily steering it towards more efficiency and transparency, much to the relief of the Limited Partners.
While we loosely agree that AI can potentially transform the way private equity functions, but this notion is a mere statement until it is broken down into relatable action points. Here’s how:
Deal Sourcing – One of the primary functions of a General Partner or an asset manager is to continuously scan the market and figure out opportunities to invest the piles of cash they probably raised in the last boom. This requires considerable amount of time and effort, and not to mention, a certain knack of maintaining relationships and levered information exchange, which doesn’t come easy. AI can help. What if a machine could scan all the financial statements of say companies belonging to a basket and figure out the shortfall between liabilities and assets and hence the need for funding? Or simpler, scan the internet for any potential investment requests posted over the internet over the last 24 hours and weed out the un-sexy ones based on the fund strategy? All it takes is a machine to get smart and start thinking.
Deal Evaluation – And here comes the tricky part, the basis of making an investment decision. Evaluating deals is where most of the efforts of the GP are involved. With AI, the models can be built quicker than they can with Excel, and not to mention more error free. The software would analyze all pros and cons of making the deal readily, with access to insurmountable amounts of data, the ways in which you could evaluate a company could be infinite. Internet is infamous for having a terrific memory, imagine if the company ever filed for a bankruptcy claim 10 years ago and nobody knows about it except the now online government archives? You could dig it up. Combining the research with an intelligent algorithm to enable triggers on certain flags taking off, we can actually see the result in several scenarios in fractions of a second. Pressing F9 each time we make a small change in the model is cumbersome, but developing a model smart enough to figure out the links to refresh every time there’s a change is so much better.
Diligence – Diligence is how you make a deal, and diligence is how you break it. Deal evaluation and diligence go hand in hand, and quite often this is rated as the most important aspect of the deal process. Hiring external consultants to do the job isn’t easy, more so when they are new and inexperienced. The established players in the market have a predefined check list, and modifications aren’t welcome. So? Imagine if there’s a system which has a query tab, as simple as the one in Google, which would be able to conduct a diligence on a company as soon as you press the return key. The system would have the power to dig out any kind of information archived anywhere in the world, which would link up to the company. Be it business, legal or financial, the AI would be smart enough to slice and dice the information and throw up the red flags. The 3 months or 4 weeks of painstaking effort of checking and correcting improper financial accounts, incomplete registrations and long ago embezzled transactions would be so easily avoided, and we would have the information that needs to be focused on, without wasting time in obtaining it.
Nobody likes talking about deals that didn’t happen. And thus, all the learning that a fund manager goes through in the process remains restricted to his mental silo, much to the disadvantage of the other managers in the same fir. What if there was an AI which could harness the thinking tanks of all the fund managers, and digitize their deal experience? That way, the learnings could be shared by a system which would warn a prospective deal if it’s heading in the same forsaken direction.
Exits – Profit has no meaning unless there’s a sale. The exit opportunities need to be constantly monitored. Need I say more? AI can definitely budge in and do the job of fixing the pieces of the jigsaw together.
For the sake of brevity, the scope of activities that AI offers in the industry is immense. Not only would faster decision making help, but there would be added benefits of increased productivity, a predictable decision making pattern, and better strategy implementation.
So what’s stopping us?
Nascent technology – The obvious challenge is the current state of AI. Even though a faster growth is promised, evident by the inflow of millions of dollars, but to develop a system which synchronizes your thoughts and converts them into action points will take time, particularly in investments. In a way, the systematic risk of the technology, given its state and the rate of development is inherently high, and the unsystematic risk of operating it specific to private equity isn’t a pleasant number either. In short, a bitter outlook. Until and unless, the basic structure of this nascent technology gets developed and widely adopted, nothing more can be said about what can be done.
Difficult learning curve – We have all dealt with the struggles of auto-correct on our phones. It takes a certain amount of time to make the dictionary get used to your unique words. All the users of the Apple watch can confirm the importance of spending this learning time with the machine in order for it to gauge the expectations and hence the output. It would take some time and a lot of painful effort in teaching the machine how to invest. Hiring people with the requisite skills looks easier.
Difficult simulation – The peculiar thing about private equity or venture capital is that it’s extremely difficult to simulate the real world environment in which the decisions are made. Not only are the quantitative market factors important, which AI can promise to take care of, but the qualitative factors particularly relationships with the transacting parties is the greatest force behind a deal. Can a machine understand the softer aspect of being a human is debatable, but would the AI on which we rely to make the decisions for us exactly act the way we do is the question worth asking.
Prediction risk – Life at best, is uncertain. Not everything can be predicted, or for lack of a better word, analyzed. Market collapses, Black Mondays, Fridays and Thursdays can’t be programmed in, and the reactions that we have in those situations are also contingent on the circumstances. Can we ensure a smooth market environment if we rely heavily on AI to do the background work? Even better, would you invest in a market where there is no arbitrage and everything is calm and predictable?
AI offers a lot of potential, without a doubt. But there’s a long process of cutting through the woods and making a track to get to the other side.
Artificial Intelligence is poised to become the future of everything and consequently, the most active arena for deal making in the next 3-5 years. AI might appear extremely simple, all you have to do is write a few intelligent algorithms for the machine to respond exactly like humans would in a given scenario, and the job is to think of all the possible scenarios and reactions, but is that the point? Why go through all the trouble of image recognition, deep machine learning and speech recognition? The answer lies in a simple Darwinian concept of relevance and independence. What AI is attempting to do is to develop a new species altogether – one which does not spill coffee while driving a car, or one which uses the internet to practically perform any task you would ask it to.
To err is human, and it’s always human. With the help of AI, the human error in cancer detection has been significantly reduced. Driverless cars are making the headlines of a soon-to-be a mass reality. Facebook and Google have been using AI to target the advertisements and get an enhanced user experience. And we’ve been able to pop up recipes on our phones, thanks to Siri. With all these positives to go for it, in the specific case of private equity and venture capital as discussed above, AI can be of an immense advantage.
While there are governments opposing the growth and development of AI systems and raising concerns over the future of humans if our own creations overpower us, but have we ever been afraid to build something just because it’s bigger than us?
Maybe, this is it.
Just as human beings have no boundaries to their thoughts and imagination, artificial intelligence lacks these limitations too, and for the better.