Understanding NLP to enhance UX in possibly all digital products!

Original article was published by Mehek Kapoor on Artificial Intelligence on Medium


Scope of NLP and industries in which it can be used

NLP has a wide scope and it can be used in any industry ranging from — education, fin-tech, social media, travel, healthcare, e-commerce and more.

Let’s see how it can be used in different industries:

NLP can be used to categorise or organise educational documents for students and teachers.

Going through books and journals and extracting meaningful information would be one job of an NLP model in education sector. Where reading so much text is almost impossible for one person, but the crux of the story can be produced using well-trained NLP models.

There are some apps that give a summarised version of books (like 12min book summaries, Blinkist etc.), and many of those apps ask for summaries from their in-house or freelance writers, who have read those books.

Those summaries sometimes include a lot of personal bias (as it has been mentioned in many app-store and Playstore comments, wherein another reader thought and understood the book in a different context), which can be avoided if machine learning is used.

For financial apps and websites, NLP can be of great help! It can help the banking officials to verify documents of customers quickly. Information can be cross-checked after a scan and discrepancies can be pointed.

Moreover, chatbots can become more reliable in future for all sorts of banking and finance-related information. Automating a lot of tedious processes in banking sector (like verifying the creditworthiness assessment of a client through their social media footprints or getting as much information about a client from online websites as possible) can also be done via NLP.

There was a story that was widely circulated in India, that people shouldn’t post pictures of their new cars, holidays, vacations, new homes, even new Apple iPhone etc. on social media (basically don’t flaunt your rich lifestyle). Why?

Because income tax department is going to analyse and use those pictures, to catch people in case of tax evasion! Tax evasion is a common practice in India, with hardly 14 million people paying income tax in the country with a population of over 1.3 billion!

UX in healthcare has just started it seems. A lot of new products are being devised and launched in the field of medicine, and it seems like it’s only the beginning of what we can imagine. NLP in healthcare can prove to be revolutionary.

I was working in a healthcare start-up in India, that aimed to digitise existing on-paper medical records of patients, and they were doing it via machine learning models. India, being a country with over a billion people, has dire needed of regulation and centralisation of medical data.

The biggest problem that users in small towns, villages or remote rural areas faced was — losing their medical records. Either to natural calamities or they just misplaced them. Many new mothers lost track of the vaccination of their child because medical records were lost.

Such problems can only be tackled if medical data of every person is available online. Using NLP, documents, prescriptions, reports, scans etc. were being uploaded to a system and important information through those reports was being extracted to build a patient-profile.

Right now, I am working in travel technology as a designer. NLP can be widely used in travel industry to extract data about every single customer from social media and other places, and present them customised deals, vacation rentals and travel packages. Not only this, NLP can also help the travel company understand more about their apps/websites by actively gathering customer feedback and generating a customer sentiment towards the brand.

I don’t think I even need to explain how NLP can change social media, because it already has to a large extent!

If you click on an advertisement on Instagram, a similar ad is available on Facebook, and whatever you search on Facebook pops up on YouTube. It’s not uncommon how one website uses the data from another website to create a customer profile at the back-end, and present them with offers and deals they can’t say no to! Hashtags on Instagram and Twitter are being used in the same way, to present ads/content to you that’s relevant to your profile.

If you watch the Netflix documentary ‘The Social Dilemma’, you can see how they show a physical model (like a voodoo doll), that is being controlled by a social media company behind the curtains, and how only customised information is provided to the user, so that he is forced to be engaged in his phone, all the time!

Machine learning doesn’t have only upsides, but it also has various downsides as well. Using sentiment analysis itself, the sentiment of audience/users can be predicted and information can be fed to them that changes the sentiment, in favour of someone/something else. Similar to that controversial Cambridge Analytica scandal, that is blamed to change the course of USA Presidential Elections in 2016.

So, with machine learning algorithms, also comes great responsibility to be fair and just.

Shopping online is only going to increase with internet becoming more accessible to upcoming/new users, generation-Z to be precise.

NLP can be utilised to provide more personal, customised and exclusive experience to each user, while online shopping. Customer service department can be transformed by changing the hassled process of pressing 1, 2 and so on, into a quicker, more informed response system that either solves customer problems or immediately transfers the call to concerned department.

Enhanced information discovery is also going to change the way feedback systems work at big brands and online retail stores. Information analysis will further help those brands to improve their shopping experience, and even introduce new integrated tech into retail like VR glasses, AR plugins, mixed reality and more.