Natural Language Processing (NLP) is one of the longest-standing areas of AI research. The idea of being able to speak to a computer and be understood, whether verbally or in writing, has been around for as long as the idea of artificial intelligence.
These days, NLP has gone far beyond being merely a better input method – we’re able to use machine learning algorithms to understand, assess, and even synthesize text and voice in unprecedented new ways. Given that marketing is heavily reliant on words to convey messages about people and products, it’s not surprising that NLP has carved out a large niche in marketing technology.
So, just what are some of the places we’re seeing NLP take off in MarTech?
If this in-depth educational content on using AI in marketing is useful for you, you can subscribe to our Enterprise AI mailing list to be alerted when we release new material.
1. Chatbots for capturing leads
It’s rare to find a website that doesn’t have a pop-up chat box on the home page offering to assist you. You can even ‘hand build’ a chatbot in Facebook Messenger to act as an autoresponder. Platforms like Drift and Intercom are typical, offering automated response platforms that can also gather information about your visitors. Currently, these chatbots tend to either come across as a bit wooden once the conversation becomes more complex, or they rely on being able to hand off to human customer support personnel when things become interesting.
The latest research now shows it’s possible to create chatbots with backstories and even the semblance of a personality by inputting just a few lines to sketch out the desired traits. This includes the ability to mirror the person interacting with the bot. Soon we will see a whole new world of realistic chatbot personalities that mirror the person being spoken to, something that has been shown to develop rapport (and will enable our bots to get more information before handing off to a human).
2. Voice search for gaining access to a wider audience
The number of people who are comfortable typing has always been a barrier to access when it comes to digital services. Voice search has become increasingly popular in recent years, from smartphones powered by Siri and Google Assistant to the advent of ‘voice-only’ speaker systems like Alexa.
Currently, 65% of 25-49 year olds speak to their smart devices at least once per day, and smart speaker shipments doubled year-on-year from Q3 2017 to Q3 2018. It’s estimated that half of the online searches will use voice by 2020, making voice an essential platform for the marketers of tomorrow.
3. Sentiment analysis for understanding customers
As NLP capabilities demonstrated significant progress during the last years, it has become possible for AI to extract the intent and sentiment behind the language. This can be used to derive the sentiment of conversations with individual customers and steer the conversation towards a conversion, as with the Vibe’s Conversational Analytics platform. It can also be used to look at the sentiment of large groups and direct group conversations, as offered by Remesh.
The latest research breakthroughs in sentiment analysis include incorporation of commonsense knowledge into a deep neural network to improve identification of aspects and sentiment polarity or using convolutional neural networks (CNNs) with gated mechanisms instead of traditional long short-term memory networks (LSTMs) for greater efficiency and better performance.
4. Automated summarization for early identification of trends
Another use case for NLP in marketing lies in the area of relevant news aggregation. The state-of-the-art text summarization approaches enable marketers to extract relevant content about their brand from online news, articles, and other data sources.
According to Stanford University’s Abigail See, NLP-based automatic summarization has moved on from simplistic extractive summarization, where content from the original source is shortened and rearranged. Increasingly, AI is able to generate summary text, creating a far more natural sounding end product. Abstractive summarization is a real game-changer, creating summaries made up of generated content on the fly from a broad range of sources.
5. AI copywriter for efficient ad generation
Struggle with creating an effective ad slogan? AI can do this job for you. The recent text generation techniques can assist advertisers in generating optimized keywords, advertising slogans, product listings and more.
For example, Alibaba has introduced an AI copywriter that undertakes much of the drudge work of creating effective product descriptions. This tool is particular popular among foreign companies that leverage this AI copywriter to create product descriptions in Chinese. There are several startups offering a similar service, Persado and Motiva are just some of the companies offering AI ad optimization. These bots aren’t likely to put anyone out of a job just yet, though they do make an excellent tool if you want to free up your writing talent to pursue more interesting tasks.
6. AI writer for efficient content generation
The idea of artificial intelligence generating convincing ‘Turing test passed’ content has been around since before Turing first posited his famous test, 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 we’re finally able to meet the reality of what those inventors dreamed of. You may recall the OpenAI case from earlier this year when a company has created a language generation model that they didn’t feel safe about sharing with the public because of risks related to the fake news generation.
However, there is still a long way to go. Commercially available systems are yet far from generating long, meaningful, and coherent texts. Even the OpenAI’s ‘dangerous’ GPT-2 sometimes gets context wrong and has been known to repeat itself.
7. Machine transcription for searchable video content
Improvements in voice recognition have led us to the point where it’s now possible to automate video and audio transcription. This opens up the opportunity to convert video and audio files to include searchable content, making these mediums more viable for improving SEO or creating human-level accurate text for an accompanying article.
YouTube already offers machine-generated captions for all of its videos. In the future, we can look forward to machine audio and video translation on the fly as well.
In short, NLP is set to continue being one of the main ‘go-to’ AI technologies for marketers. It’s also going to be part of the leading edge of MarTech evolution for the foreseeable future.
These technologies will grow closer to what a human could achieve rapidly over the next few years, and in many cases will drastically change the face of marketing departments and roles. Like with all the latest AI developments, it’s important for marketers to learn how to get the most out of these tools if they want to keep themselves and their skillset relevant as we head into the future.
Enjoy this article? Sign up for more updates on using AI in marketing.
We’ll let you know when we release more technical education.