According to the recent Gartner report, “Top 10 Data and Analytics Technology Trends That Will Change Your Business,” Natural Language Processing (NLP) will have a transformative effect on data analytics. By removing the need to program queries into an analytics tool, NLP expands the potential user pool, along with use cases, for advanced analytics. For the financial field, to choose one of many potential areas of impact, NLP and conversational analytics promise to be revolutionary.
What is Natural Language Processing, or NLP?
NLP has long been a goal in computing, ever since the robot in Lost in Space said, “That does not compute” or the HAL 9000 uttered the infamous words, “Sorry, Dave, I can’t do that…” Getting to a viable and economical NLP capability took some time, however. There were many false starts along the way. Today, though, NLP is an established branch of Artificial Intelligence (AI) that enables interactions between computers and human beings using natural language.
NLP enables a computer to understand spoken or written human language. For example, people who manage stock portfolios are constantly assessing the financial risks faced by the companies whose stocks they own. A wide range of factors can affect financial risk, including exposure to bad debt, interest rate exposure, contingent liabilities, industry-specific and macroeconomic issues and on and on.
To conduct a financial risk assessment for a single company is an extended, labor-intensive process. As a result, portfolio managers only perform a full analysis periodically. With NLP, however, an analytics solution can constantly “read” through enormous volumes of data related to the company in question. It can plow through huge repositories of documents and identify topics that are relevant to the company in question. Through this process, an NLP-enable analytics solution can do more in a day than a person can do in a year—improving the risk management aspects of portfolio management at the same time.
This is a lot harder than it looks. To work, the system has to develop a taxonomy or knowledge. It needs to “learn” what risk assessments are about, and so forth. This requires a great deal of Machine Learning (ML) and language rules.
2019 Market Guide for Data and Analytics by Gartner
Gartner sees a significant future for NLP in analytics. They predict that 50% of analytical queries will be generated via search, voice or NLP (or automatically generated) by 2020. The further foresee that NLP and conversational analytics will drive analytics and business intelligence adoption from 35% of employees up to over 50% by 2021. According to the research, NLP will enable whole new classes of users, such as front office workers, to take advantages of analytics tools.
How NLP and Conversational Analytics Will Change Everything
NLP and conversational analytics promise profound changes in the business world, especially in finance. Finance is very competitive. Success depends on gaining an edge over competitors. This could be an advantage in trading strategy, a better understanding of the customer’s needs, more efficient operations and so forth.
Data is invariably the key to unlocking such advantages. One of the difficulties, though, has been a lack of data literacy in the financial firm. Using data analytics tools is challenging, even with recent advances in usability. With NLP, though, non-specialists can easily query databases as well as other, less organized collections of information.
NLP and conversational analytics potentially revolutionize the financial field by enhancing the firm’s knowledge of the customer. Customers are constantly providing data about how they really feel about the firm. The challenge is finding that data and interpreting it correctly. It’s not sitting in neat database rows and columns. Customer knowledge is hidden in social media posts, PDFs of account applications, customer service helpline call records and so forth. An NLP-enabled analytics tool can get at these elusive bits of customer data and help your firm benefit from the resulting customer insights.
These analytical tools can help you find a “needle in a haystack.” By combining NLP with augmented analytics, e.g. automatic insight generation, users can quickly discover difficult-to-find information. In finance, this capability might apply to internal audit and fraud detection. Finding a suspicious transaction among billions of nearly identical transactions is monumentally difficult. However, with NLP and conversational analytics, the toolset can compare and correlate dozens of data streams that inform the transaction analysis process, resulting in finding that critical “needle in a haystack,” or, as one fraud analyst once put it, “a needle in a pile of needles.”
Understanding human language is challenging in a narrow professional context. NLP interfaces address this issue by embodying industry- or domain-specific language. You might want to research stocks that have been subject to a “short squeeze,” to use an example of a somewhat arcane financial term. By teaching the tool what a “short squeeze” is all about, however, it can then go hunting for situations where short selling made a stock price go up unexpectedly.
Want to learn more about NLP and what it means for your BI and data analytics program? We can help you assess the potential impact of NLP on your business.
Written by Tripp Smith
Data analytics thought leadership spanning entrepreneurial ventures and Fortune 500 companies centered around big data, technology vision, strategy, and product development. Chief Technology Officer at Clarity Insights focused on technology vision, product development, business development, partner alliances, coaching, and mentoring for the largest consultancy in the US focused solely on data analytics.
Topics: Natural Language Processing