With 40,000 listed stocks and more than 100,000 investment funds, how can financial services firms manually gain insight into investment data? They can’t, which leads to a great deal of investor time and manual effort being consumed by research, creating latency that can slow portfolio decision cycles, resulting in lost upside and higher risk exposure.
Part of the problem is the amount of unstructured investment data that needs to be evaluated. Ninety percent of all data is unstructured and less than 5% of it is analyzed. Investors struggle to interpret unstructured data within emails, instant messages, conference calls, social media posts and content — and 30% of financial firms don’t have a formal process around unstructured data management.
AI as a Financial Services Game Changer
More financial services firms are turning to AI to streamline operations to do more, faster, and using analytics behind their decision-making. Natural language understanding (NLU), an AI technology, provides a deep understanding of unstructured text data associated with investment research. This includes the ability to extract, analyze and evaluate investment and risk indicators hidden within financial documents and disclosures, so investors can make informed evaluations and capitalize on opportunities sooner.
Critical to NLU is the approach used to analyze the data.
Machine learning (ML) relies on training and pattern detection to identify information. This results in a black box approach where a tremendous amount of data is necessary to train the algorithm, but there is no insight into the algorithms used to interpret the data. Furthermore, because a ML-only approach does not actually understand language, the context and ambiguity inherent in text data is difficult for even the most sophisticated systems. Symbolic AI is a rules-based approach that uses embedded knowledge in a way that is similar to how humans learn. Because the rules are written by humans, they can be easily re-written to address any issues. This distinction is important when it comes to financial services because the algorithms behind investment decisions must be explainable and transparent.
Hybrid NL takes a best of both worlds approach that brings together a human-like understanding of language with the data processing capabilities of machine learning to enable deep understanding and insight at scale, with outcomes that are explainable.
Here are six ways financial services firms can streamline operations and improve outcomes for their customers with NLU platforms using a Hybrid NL approach:
1. Find Risk in Financial Disclosure Documents
Corporate disclosures protect companies from liability should their financial forecasts miss their mark due to changing economic conditions. They also show company relationships to avoid conflicts of interest. Yet disclosure documents are written by lawyers with text that’s often difficult to read and understand. An NLU platform can extract key data from disclosure documents to be analyzed to reveal future trends, potential conflicts of interest and warning-like statements. In doing so, investors can reduce the manual effort and time needed to evaluate the most vital information from these disclosures.
2. Discern Earnings Call Sentiment
Most of the 3,000+ U.S. publicly traded companies listed on major stock exchanges hold earnings calls every quarter. Valuable investment signals can be found within these calls, including chief executive comments on business conditions or profit expectations that move a company’s share price. Q&A segments can capture candid remarks revealing sentiment and intent, while tone can indicate optimism or disappointment. Financial services firms can use NLU platforms to evaluate and weigh investment signals from earnings call transcripts. For example, it can decipher risk based on a decrease in revenue, a hike in the cost of goods sold, a change in a management team or a drop in earnings per share.
3. Scale Fund Prospectus Evaluation
A fund prospectus provides investors with information about stocks, bonds and mutual funds, including fees, strategy, ownership structure, performance history and risk profile. Also included is information about environmental, social and governance-related (ESG) risks and opportunities that may affect an investment. An NLU platform can be used to quickly and accurately reveal insights from prospectus data, including investment strategy, fees and fund performance over time. When you consider that it can take the average person more than 16 minutes to read a 10-page document, NLU can automatically analyze text data in prospectuses at scale with less manual effort.
4. Find Actionable Annual Report Data
The average annual report is the equivalent length of a 240-page novel – and 80% of the content is text that includes complex verbiage. Hidden within this text is valuable information that can provide further context and insight into a company’s assets, balance sheet income statements, earnings, cash flow and forward-looking executive outlooks. Financial services firms can use NLU platforms to mine annual report data, extracting and analyzing language data to fuel insight that can help investors rule opportunities in or out faster, with less chance of missing risk signals due to fatigue or misleading language.
5. Aggregate Company and Sector News
Investors are challenged to stay informed of company and sector news because of the constant flow of information from multiple sources. NLU can be used to extract metadata from a wide range of structured and unstructured data formats, pulling relevant information that can be summarized across sources. Triggered alerts into entities, topics and events can provide investors with targeted information in real time. This ability to extract and filter information from the news “noise” helps investors more clearly identify cross-cutting trends opportunities within company and sector announcements.
6. Extract Social Media Trends
Social media has the power to move the needle on stock highs and lows. For example, when Elon Musk tweeted that Tesla would no longer accept bitcoin as payment, the price of bitcoin dropped about 15%. Social media offers a real-time view into unfiltered attitudes, opinions and outlooks towards brands. Financial services firms can use NLU platforms to derive data from social content with triggered alerts across social channels. This brings to the surface key social media trends that can potentially impact specific stocks or industry sectors.
Expert.ai for Streamlined Financial Services
Hybrid NL platforms provide a deep, human-like understanding of unstructured language data and turn it into knowledge and insights so investors can make faster, more consistent decisions.
Expert.ai delivers NLU to support key financial processes.
We help financial services firms streamline their information-intensive operations and realize the importance of language-based data. Our NLU uses Hybrid NL to extract and analyze critical information from financial records and regulatory disclosures so that firms have full transparency into how the data was interpreted.
With expert.ai, investors can mine for text-based insights into financial documents faster and with fewer resources. In doing so, they can speed service delivery and deliver guidance and financial advice with more real-time intelligence and analytics.
