Head Office Singapore: 31 Rochester Drive,Level 24,Singapore,138637
+65 6808 8760
India Office: 880,
Adarsh Nagar,
Jogesheri West,
Mumbai,
400053
Head Office Singapore: 31 Rochester Drive,Level 24,Singapore,138637
+65 6808 8760
India Office: 880,
Adarsh Nagar,
Jogesheri West,
Mumbai,
400053
Influence of Natural Language Processing (NLP) on Marketing and Market Research

Influence of Natural Language Processing (NLP) on Marketing and Market Research

One of the longest-standing fields of AI study is natural language processing (NLP). For as long as the concept of artificial intelligence has been around, the concept of being able to talk to a machine and to be understood, whether verbally or in writing.

NLP has gone far beyond being merely a better input method these days – we can use machine learning algorithms to understand, evaluate, and even synthesize text and voice in new ways that have never been before seen before. Because marketing relies heavily on words to convey messages about people and products, it is not surprising that NLP has sculpted a large marketing technology niche.

So, just what are some of the positions we see taking off at MarTech from NLP?

CHATBOTS FOR CAPTURING LEADS

Finding a website that does not have a pop-up chat box on the homepage offering to assist you is rare. You can even 'hand-build' a Facebook Messenger chatbot to act as an autoresponder. Systems such as Drift and Intercom are traditional and provide automated response systems that can collect information about the visitors as well. These chatbots currently tend to either come across as a bit of wood once the conversation becomes more complex, or they rely on being able to hand it over to human customer support staff when things get interesting.

The new work now shows that it is possible to create history chatbots and even the appearance of a personality by inputting only a few lines to draw the desired traits. This includes being able to mirror the person who interacts with the bot. Soon we'll see a whole new world of realistic chatbot personalities that mirror the person being spoken to, something that has been shown to develop relationships (and will allow our bots to get more information before handing it over to a human being).

Conversational AI with a persona

Conversational AI with a persona

VOICE SEARCH FOR GAINING ACCESS TO A WIDER AUDIENCE

In digital services, the number of people who are comfortable typing has always been a barrier to access. In recent years, voice search has become increasingly popular, from smartphones powered by Siri and Google Assistant to the advent of 'voice-only' speaker systems such as Alexa.

Currently, 65 percent of 25-49-year olds are talking to their smart devices at least once a day, and smart speaker shipments have doubled year-on-year from Q3 2018 through Q3 2019. Half of online searches are estimated to use voice by 2020 making voice an essential platform for tomorrow's marketers.

SENTIMENT ANALYSIS FOR UNDERSTANDING CUSTOMERS

As NLP capabilities have made considerable improvement over the past few years, AI has been able to extract the meaning and emotion behind the language. As with the Vibe's Conversational Analytics platform, this can be used to derive the feeling of conversation with individual clients and to steer the conversation towards a conversion. As provided by Remesh, it can also be used to look at the opinion of large groups and guide group conversations.

The latest research breakthroughs in sentiment analysis include integrating common-sense information into a deep neural network to enhance the recognition of aspects and feeling polarity, or using gated systems with convolutionary neural networks (CNNs) instead of conventional short-term memory networks (LSTMs) for greater efficiency and performance.

SenticNet semantic network

SenticNet semantic network

AUTOMATED SUMMARIZATION FOR EARLY IDENTIFICATION OF TRENDS

Another marketing use case for NLP lies in the field of related news aggregation. The state-of-the-art approaches to text summarization allow marketers to collect meaningful information about their brand from online news, articles and other sources of data.

According to Abigail See at Stanford University, automatic summarization based on the NLP has moved on from simplistic extractive summarization, where content from the original source is shortened and rearranged. AI is increasingly able to produce summary text, generating a sounding end product which is much more realistic. Abstractive summary is a real game-changer, generating summaries from a wide variety of sources, composed of created content on the fly.

  Examples of highly abstractive summaries with novel words denoted in blue

  Examples of highly abstractive summaries with novel words denoted in blue

AI COPYWRITER FOR EFFICIENT AD GENERATION

Trying to fight with the development of a successful ad slogan? AI will do the job for you. Recent techniques in the generation of text can help marketers produce customized keywords, promotional slogans, product listings and more.

Alibaba, for example, has introduced an AI copywriter who does much of the drudge work of creating effective product descriptions. This tool is especially common among international companies that use this AI copywriter to construct descriptions of the products in Chinese. Many start-ups offer a similar service, Persado and Motiva are only a couple of the companies that provide AI ad optimization. These bots are unlikely to put anyone out of a job just yet, although they do make an excellent tool if you want to free up your writing talent to pursue more interesting tasks.

AI WRITER FOR EFFICIENT CONTENT GENERATION

Since Turing first published his famous test, there has been the concept of artificial intelligence producing compelling 'Turing test passed' material, as can be seen by the popularity of automata during the age of enlightenment, including one that is a programmable writer

250 years later, and at last we can meet the reality of what those inventors dreamed of. You may recall the case of OpenAI from earlier this year when a company created a model of language generation which they did not feel safe about sharing with the public because of the risks associated with the fake news generation.

There's still a long way to go however. Still far from producing long, meaningful and coherent texts, commercially available systems are. Even the 'dangerous' GPT-2 of OpenAI sometimes gets the context wrong and is known to repeat itself.

Sample of AI-generated story

Sample of AI-generated story

Concluding Remarks

In short, NLP is expected to continue to be one of the marketers' main 'go-to' AI technologies. It will also for the near future be a part of the leading edge of MarTech evolution.

Such innovations will evolve closer to what a person might do rapidly in the next few years, and in many instances the marketing departments and positions will radically change face. As with all the latest developments in AI, it is important for marketers to learn how to get the most out of these tools if they want to be relevant to themselves and their skillset as we move into the future.