Language AI adoption continues to rise as organizations across sectors pursue digital transformation in a context of pandemic disruption. Budgets reflect the growing interest. A global survey conducted by John Snow Labs shows NLP budgets increased: 60% of respondents say their budget for natural language processing — the official term for language AI — increased by at least 10%, and 33% say their budget increased by 30%.
General research on AI shows companies starting to see ROI, meaning that those budgets are starting to generate value. Companies that reap above average returns are now sharing best practices, one of which is pursuing clearly defined use cases. Here, Persado highlights three areas where language AI is delivering value: better customer experiences, more usable human insights, and stronger decision making.
Before exploring these specific language AI use cases, it’s worth revisiting what the core capabilities of natural language processing are.
Natural language processing is the field of AI concerned with recognizing and understanding the words humans use when they speak to each other, and producing recognizable responses. The majority of data businesses have on customers is in the form of language — think of the conversations customers have with service reps, or with sales people, or through email, presenting an untapped opportunity for optimization
But businesses couldn’t efficiently analyze that data until recently, because language data is unstructured, meaning that a machine can’t automatically interpret it. Businesses need models that perform a translation function to tell computers how to understand different aspects of language — such as what a word is and what it means, as well as structures of speech like syntax.
Language AI provides that function. And depending on the application, a machine can understand different things from language depending on the capabilities of the AI. Example capabilities include:
Those capabilities reveal clear ways businesses could benefit from language AI. A technology company, for example, could more quickly identify an outage or a product bug by clustering annoyed calls about the same issue; or a healthcare firm could more efficiently anonymize patient records before analyzing trends in patient visits. Or, as is the case with Persado’s customers, a business can find the ideal way words, images, and structure for any variety of campaigns in order to motivate actions that both meet customer needs and produce business value.
The past ten years have seen not only growth in the number of models available to businesses (for example, GPT-3), but also an explosion in experiments to benefit the business. The ones gaining tracking and producing benefits are serving to improve customer experiences, gather insights, and drive more effective and efficient decision making.
PwC’s AI Predictions 2021 report found that “creating better customer experiences” was where the highest percentage of leaders (86%) reported seeing benefits from AI. One specific use of language AI toward that goal is conversational AI. Many people have “conversational AI” interactions every day when they ask Siri a question. Businesses benefit for example in using chatbots to help with customer service. Chatbots are AI-powered language engines trained to deal with common, often straightforward, customer service concerns and save human agent time for more complex issues.
Accenture predicted in 2018 that by now, 15% of customer service interactions would be handled by AI. Gartner notes that growth varies company to company depending on how a firm designs its channels, since customers often default to human interaction. Better design and testing of “contact us” web pages could help the one-in-four companies that are already piloting chatbot technology drive customers to the chatbot instead.
Overall, telecom is one of the most active adopters of language AI, according to McKinsey research, because of the benefits of using chatbots for customer service.
Persado saw first hand the power of language AI to guide customers to self-service solutions when a customer asked us to craft an IVR script to steer callers away from the call center queue and toward self-service channels. That single campaign resulted in thousands in call center savings.
Language AI is increasingly integrated into data analytics processes to help clean and organize data and run analyses. For an example of the latter, customer data platforms — a workhorse in the marketing technology stack — are leveraging language AI to run customer sentiment analysis, among other use cases.
Sentiment analysis using language AI interprets the emotional context of an interaction so that businesses can identify and track trends in customer attitudes related to the business, a product or service, or a circumstance. For example, consumer technology firms might track customer sentiment related to a product launch, or investment management firms might use customer sentiment to predict investment behavior.
Sentiment analysis isn’t just for internal communications, either. Organizations can leverage it to analyze conversations over social media or in online forums to capture insights. The benefits include gaining early warning of changes in customer attitudes or predicting new needs to guide product development. Internally, businesses can use sentiment analysis to gauge employee engagement and satisfaction — a key issue given the current mass exodus of workers.
The ability to understand how customers feel and communicate with them in a targeted way is a hallmark of targeted-to-personalized marketing. In a demonstration of its power to engage customers, a Fortune 50 bank worked with Persado to craft differentiated messages for credit card reward customers. The emotional tone of the message that motivated customers who typically sought travel rewards was different from those who redeemed for cash rewards. By fine-tuning the message to reflect the right emotion, Persado increased campaign performance above the control by 60%, for travel redeemers, and 44% for cash redeemers.
Most business leaders agree that better and more accurate information should lead to faster, and higher-quality decisions. Language AI helps through its ability to interpret large quantities of formerly uninterpretable data captured through verbal and written interactions.
One industry testing that theory is healthcare. Thirty percent of healthcare firms surveyed by McKinsey have adopted language AI to improve operations. Some use cases include scanning patient records to spot patterns that point to potential outbreaks of COVID-19 or flu. Another involves voice-to-text capabilities that allow doctors to input content into patient records verbally during visits, when the patient can correct or qualify in real-time so that there is a more accurate record.
Persado’s work with the bank Emirates NBD offers an example of AI-facilitated decision making. The bank worked with Persado to quickly run a large number of language experiments that captured learnings about the Emirates NBD customers and the concepts, emotions, words, and language elements that motivated them. With sufficient audience insights, Persado was able to leverage the decisioning power embedded in its language AI platform to generate predictive content that delivered the right message without experimentation. This significantly accelerated the time it took to deploy the best content into the market.
AI investments are expected to continue growing in the coming years, with language AI among the prime areas delivering value. For businesses looking to achieve clear short term value, improving customer experiences, capturing more usable insights and enabling better decision making present clear opportunities.