The job of a sell-side analyst can be broken down into two basic parts: first, gathering relevant data, and second, producing insights based on that data. The latter function is largely up to the intelligence and experience of the analyst. The first, however, has long been ripe for technological intervention.
Searching through endless regulatory filings, company presentations, earnings call transcripts, news releases, and other research is an interminable and eye-glazing process. Electronifying those documents has allowed analysts to perform keyword searches, although they still had to search each document for every keyword variation individually.
In 2010, former Morgan Stanley analyst Jack Kokko decided more could be done. He teamed up with technologist and one-time Wharton School of Business classmate Raj Neervannan to create AlphaSense, a search engine for financial analysts that does for research documents what Google does for webpages. Using semantic technology, AlphaSense enables thematic search of millions of documents instantly, leveraging linguistic search algorithms tailored for investment research. It semantically indexes unstructured text documents sentence by sentence, while utilizing structured data for further filtering, and produces instant results on any theme.
“One of the most unique features of our search technology is its ability to capture variations in language across companies and over time, so our users can focus on searching for themes rather than simple keywords,” says Kokko, now AlphaSense CEO. “For example, our taxonomies capture over 150 synonyms for ‘growth,’ such as ‘rising’ or ‘gaining traction,’ while filtering out comments about ‘giving rise to’ or some 100 other anti-patterns. While language for many common topics like ‘growth’ or ‘revenue’ changes very little over time, ‘Fed taper’ was hardly a recognized term until recently, and in the past year, it became one of the industry’s most talked about themes. Our algorithms and taxonomies continually evolve to capture the language changes that occur naturally as the markets and macroeconomic conditions evolve.”
“When everyone has access to the same information, the ability to use AlphaSense to quickly cut through the noise and connect the dots can mean the difference between acting on a unique investment idea and going with the flow, along with everyone else.” —Jack Kokko, AlphaSense
Users can search with pre-built thematic filters as well as their own keywords. Searches can combine unstructured thematic text search queries with structured data filters such as ticker, watchlist, Global Industry Classification Standard (GICS), market capitalization, document source, and date range. New disclosures are tracked automatically, 24/7, and users can be notified through instant email alerts. “When everyone has access to the same information, the ability to use AlphaSense to quickly cut through the noise and connect the dots can mean the difference between acting on a unique investment idea and going with the flow, along with everyone else,” Kokko says.