AlphaSense – Artificial Intelligence for Financial Data
We know that we can use artificial intelligence (AI) for trading stocks, and some hedge funds are making a killing in this space. In order to feed these algorithms, there are generally two types of financial data you can use; fundamental and market data. If you think about all the data related to a stock that describes the way it trades, you’re thinking about market data. These types of data sets can become very large, very quickly when you consider just the following market data sets:
- Price change for every second of every day
- Number and size of trades for every day
- Each bid/ask changes through the day
Using this data to make trades would be referred to as “technical trading” because it ignores the fundamentals behind stocks. People that do technical trading are often thought of as speculators by investors who invest based on fundamental data, things like profitability, revenues, valuations, etc. All of the fundamental data for every publicly traded stock on a major exchange is contained in regulatory flings which most people don’t see, but that is accessible using online tools. These are the tools we use to expose over-the-counter (OTC) companies that are misleading investors. The only problem we’ve found is that it takes a lot of time to dig through the data. Here’s what the interface for the SEC online filing system looks like:
It’s not very user friendly, but the fact is that every single accounting metric that you can think of can be derived from the filings of publicly traded stocks like the ones seen above. One company that made the CB Insights AI 100 list is AlphaSense, a startup that is using artificial intelligence to make all of the world’s fundamental data accessible at your finger tips.
Founded in 2010, San Francisco startup AlphaSense has taken in $35 million in funding so far with the lion’s share of that funding ($33 million) taken in during March of 2016. With a product out since 2010, this means AlphaSense is getting ready to scale. The company is using natural language processing algorithms (a branch of artificial intelligence) to comb through not just the SEC online regulatory filings but also more than 1,000 sell-side research sources.
If you’re fortunate enough to not work in finance, what sell-side means is this. Essentially you have thousands of pundits who get interviewed on shows like CNBC’s Squak Box and wax poetic about which shares they like/dislike. As it turns out, their research methods are akin to throwing darts at a dart board. Actually even worse, because the active fund managers who buy this research only outperform their benchmarks 20% of the time. This collective group of tools are what we refer to as the “sell-side”.
Whether that research is accurate or not, it contains immense amounts of useful information that these analysts spend 18 hours a day putting together in hopes that they’ll make VP someday and be able to afford a closet full of Ferragamo ties. The AlphaSense platform indexes all this information (makes it easily searchable in other words) and also uses some clever tricks to ensure that all that financial terminology out there can be understood by a computer:
Want to quickly query every issue of the Wall Street Journal, Barron’s, or even the Financial Times? Here are some examples of the types of the natural language queries you can use on the AlphaSense platform:
- Who is spending capital to increase their production?
- What are all the new risk factors in the latest 10-K?
- Who mentions my companies in their earnings’ calls?
- What has management said about capex recently?
Astute readers will notice that the last two bullet points above can’t be taken from regulatory filings or media articles necessarily. You’d be exactly right. AlphaSense also combs data sources such as conference call transcripts, investor relations presentations, and even your own content which you can upload. You can search across millions of documents with a few clicks and get your answer in seconds for 35,000+ global companies. They’ve won all kinds of awards and count over 500 customers including 60% of the world’s hedge funds. Worried about security? AlphaSense uses Amazon Web Services to host their cloud based solution which means they can stop worrying about security and just focus on improving those algorithms.
As finance professionals ourselves, we became giddy as schoolchildren at the notion of a platform that can do what Alphasense can. The amazing thing about artificial intelligence is that while we’re debating whether or not it’s ripe for disruption, it’s actually already creating disruption right underneath our noses. We used to rely on hundreds of IIT graduates in Mumbai to sift through all the financial data for the 10s of thousands of publicly traded stocks. Now, we can simply use artificial intelligence to get those same numbers and it never makes a mistake. All the financial data vendors (there are 100s) need to pay attention to what direction Alphasense is heading. Maybe it will help explain some of those unexpected cancels you’re going to face next year.
Paying too much in transaction fees? Check out a brokerage firm called Zacks Trade that's offering $1 trades on U.S. stocks and options until 2020. After that, you'll pay just $3 a trade or a penny a share, whichever is greater. You can also trade on 91 stock exchanges in 19 foreign countries. Click here for $1 trades on U.S. stocks and options until 2020.