Natural language processing (NLP) is changing the way humans interact with machines in ways that were unthinkable a decade ago. Thanks to huge advances in machine learning, driven and supported by ever-faster computer processing, we are increasingly using NLP tools such as Amazon Alexa, Siri, Google assistant, Cortana, Bixby, and interacting with chatbots from our service providers.
Using these devices/services we can get answers to questions from the trivial – what’s the weather forecast for London – to the complex – navigating a Choose Your Own Adventure interactive story or finding out how your investment portfolio is performing.
IBM’s Project Debater uses AI technology to argue with humans on complex topics and represents the potential for NLP in all areas of society. Over time, NLP will be programmed to understand more complex elements of the human language such as humour, sarcasm, satire, irony and cynicism, according to Analytics Insight.
“NLP has taken off in different areas, one being chatbots,” says Jim Lewis (pictured), Director, Solutions Strategy & Development, Investment Manager Services Division, SEI. “The ability to transition from bot to human then back again is a really interesting development that I don’t think people have fully grasped. It tells you how pervasive NLP technology is becoming, as well as highlighting the increase in quality.” In fact, California recently passed a law stating that users need to be informed whether they are communicating with a bot or human online.
“The number of Alexa-enabled devices shows that this isn’t going anywhere. Google Home, Amazon Alexa etc., are showing there are things they can do that we didn’t think possible just a year or two ago. Amazon has even rolled out Alexa for Business, which show it’s inevitable that things are only going to progress further and at a speed we have not experienced previously.”
A report by MarketsandMarkets forecasts that the NLP market could grow from USD7.63 billion in 2016 to USD16.07 billion by 2021. Growth will be driven by demand for enhanced customer experiences and the broader application of NLP technology in healthcare, web and cloud-based business applications.
As an industry, financial services has not fully embraced the potential that NLP technology could offer. However, SEI and others are engaged in proofs of concept to explore how the twin forces of NLP and machine learning might improve the way that employees interact with machines, and in so doing, improve work efficiency and the ability to develop fresh analytical insights to share with clients.
NLP – A new dimension
“It is the conversational aspect to NLP that starts to get really interesting; the machine learning piece. You ask a question and not only do you get a response, you may get more information than you originally asked for, which you can query further, and get even more granular detail.”
“This is creating a more natural way for how we communicate with chatbots, mimicking the very same way we communicate with other human beings,” comments Lewis.
People are now using chatbots as their virtual assistants, asking questions in real time while performing everyday tasks just as they might ask a colleague sitting nearby.
This has the potential for humans to overcome the need to perform a lot of repetitive tasks and operational demands that would ordinarily require disparate systems, multiple log-ins etc., and take up a lot of time. What historically would have taken multiple people on multiple systems could soon be facilitated by a single person and a bot in the very near future. As we all know, workplace efficiency is a win for the company and its stakeholders as well as the folks performing the daily tasks; most employees want to add value to a process not perform repetitive or inefficient tasks.
“I think it is going to take the financial services industry to a whole new dimension as folks start to figure out how best to utilise this NLP technology,” asserts Lewis. “Certainly, that’s our expectation at SEI.”
“At its best case, it will allow us, as humans, to focus on higher value-added activities and avoid having to perform mundane tasks. It will free up time to do more analysis, to find patterns, and come up with more tangible information (for clients) rather than push the same buttons to complete a task.”
An analytical continuum
The suggestion here is that combining AI tools with NLP applications could be done to perform an initial analysis of data, before any human ever looks at it. At the end of the process, one could instruct the NLP application/interface to share the results of that initial analysis in a report format or spreadsheet, at which point the individual could perform a deeper dive into the results.
This could apply to healthcare diagnostics and e-commerce just as much as financial services.
“I see it as a continuum,” says Lewis. “Machine learning is great but it doesn’t solve everything. You need a thesis to work from. AI and robotics could be used for basic processing tasks, or get things to a place where the human can do things much more quickly, leaving them time to decide where to look for patterns or value add opportunities. Over the top of that, NLP could play an important role in the orchestration of events. “
“Think of it a bit like the Microsoft approach. There are multiple ways you can perform a single task in Excel; quick keys, functions etc. NLP gives you that same flexibility by allowing for a typed in request via a chat bot, a spoken request via a voice assistant, or a traditional systematic trigger from a process or application. This could provide a bridge to how we use AI and robotic process automation. It will keep morphing as companies move across the technology curve.”
SEI is in the early stages of exploring how NLP might enhance the way its staff interact with clients to improve the overall level of engagement. It is currently developing a bot framework to expand what Lewis refers to as the ‘trigger and the action’.
“You need some way to trigger an event and some way of delivering it. What channel do you trigger through? Alexa? Google Home? A chatbot? Facebook Messenger? IoT device? We’ve looked at all these and more.”
“The bot framework then performs an action based on the parameters the user sends in the request i.e. today, the user might ask Alexa to run a specific performance report and send it to someone’s email account” says Lewis. “Then tomorrow, the same user could use IVR (Interactive Voice Response) via their cell phone to send the same report to an FTP site.”
Traditionally, performing tasks has always been a very linear process. The individual logs in to a system, clicks a few buttons, then generates a report (or any other form of output). It has always been a step removed from true interaction with the machine.
NLP technology is changing that and the potential applications to business are huge.
“Ultimately, what we want is for the end user to have multiple paths to get to an end result quickly, accurately, and conveniently. This is greatly different than the current solution of the machine dictating the terms of how the task needs to be completed,” concludes Lewis. “Going forward, personalisation is going to be integral to everything we do.”