We recently published a briefing paper with the Harvard Business Review about how artificial intelligence (AI) can turn investment research into competitive advantage. In the first post of a series summarizing the findings, we discussed some of the challenges analysts face due to information overload.
Technology may have contributed to this overload, but it can also help to manage it. So in this post, we’re going to explore how advances in AI, machine learning, and natural language processing have led to the development of tools which analysts can use to overcome these challenges and supercharge the hunt for alpha. While none of these tools replace humans, they augment the decision-making process and ultimately help to build stronger investment cases.
Profit and margin are two key pieces of information when researching a company. But how many other terms can be used to describe the same thing? Profit has 45 synonyms according to thesaurus.com, and margin has 30. Different meanings typically cause programmes designed to do investment research to return a lot of useless information from scanned documents. However, semantic analysis algorithms understand sentences by breaking down their structure and recognizing combinations of words, making sure analysts only receive the most relevant results.
Analysts can use a process called topic modelling to uncover major themes within a document or a set of documents. Algorithms can be trained to scan pages of text and detect patterns or recurring words to reveal the topics that come up most frequently.
AI tools can crunch numerical data as well as textual data. They automate the tasks of identifying trends hidden in the numbers and comparing a company’s results with previous financial years or their competitors. Eliminating the need to copy and paste figures from one spreadsheet to another and having to ask a colleague to double-check their work for accuracy, means analysts can focus their efforts on processing the data and uncovering investment opportunities they might have otherwise missed.
Certain employees are critical to a company’s success. While the passing of Steve Jobs was widely mourned, the departure of Apple’s Chief Design Officer Jony Ives didn’t make as many headlines, even though he played an intricate role in the design of the iPhone (not to mention the company’s other products). Analysts can now keep track of executives and employees as they move from one job to another using named entity recognition, which takes into account variations such as middle initials or misspellings.
Market sentiment has a major influence on share prices. Emotions like overconfidence or fear lead to inefficiencies which provide a valuable source of alpha. That’s why many analysts are keen to capture the mood of investors. Sentiment analysis tracks what executives are saying about a company, which can then be compared with statements from competitors or commentary by other analysts in the financial media or on social media.
In the third and final blog post in this series, we’ll explore how AI can improve workflows and some best practices for putting these tools into action. Alternatively, you can download the full briefing paper here.