A market research firm concluded that the global market for voice-assisted speakers grew 187% in the second quarter of 2018. Cortana, Alexa, Echo, HomePod, and other variants are popping up in living rooms everywhere. Increasingly, we are navigating the digital world through voice commands. And this presents an opportunity for mar-tech, if we can make sense of the data stream.
When determining intent from a marketing perspective, the relationship argument cliché “it’s not what you said, it’s the way you said it” may apply. Companies have been able to extract revealing insights just from the patterns contained in our voices — prosody, pitch, volume, intonation, and pacing. That paralinguistic information sometimes reveals our true intent. And the ability to accurately discern consumer intent could substantially assist with the cost-effectiveness of marketing efforts.
Analyzing Voices and Emotion
By utilizing the latest computer vision and speech processing technologies, market research firms can gauge consumers’ emotional responses to proposed or existing campaigns. But affective computing technologies can also be deployed proactively, in order to refine digital experiences, support unique apps, improve the relevancy of recommendations, and influence the buyer’s journey.
One company, Affectiva, was originally spun out of MIT Media Lab and has since worked with major brands to achieve these feats. Its SDK allows developers to apply emotional intelligence to virtually anything. Devices equipped with emotional intelligence would be more in-sync with their users and could better understand intent.
It isn’t easy. High-performing algorithms in this field need to pay attention to changes in paralinguistic information as well as the larger context. Additionally, deep learning architectures rely on large datasets for training. Therefore, affective computing companies need to develop large emotion data repositories. But that investment pays off.
Each person’s voice contains a wealth of information. So much so that there are even possibilities for diagnostics. The Israeli company Beyond Verbal is developing vocal biomarker prediction algorithms to assist in healthcare screenings. In a previous interview with DMN, the company’s chief science officer told me that vocal analytics can also be used to fine-tune interactions between bots and customers.
Intent Marketing and AI: The Big Picture
Marketers can potentially uncover a consumer’s true intent through an AI-powered analysis of context, keywords, tone, loudness, tempo, and quality. If they’re consistently effective at this undertaking, marketing will be forever changed.
Intent marketing represents a fundamental shift in perspective. Instead of focusing on a predetermined brand message and disseminating it as widely as possible, marketers are now trying to discern the customer’s goal. What is their underlying motivation? Are they interested in a broad topic, or intending to purchase a specific type of product? The difference between intent and interest can be subtle and difficult to discern. But once that code is cracked, marketing dollars can be spent more efficiently. The answers lie hidden in the data. This is a job for machines, not people.
As AI gets better, it could get inside our heads better. The public often thinks about AI in a far-reaching way. Will artificial general intelligence (AGI) replicate our consciousness, or transcend it, or threaten our very existence? The mar-tech industry is, by and large, sidestepping the philosophical and going straight for the pragmatic. Companies want to understand AI so that AI can provide them with a specific, actionable understanding of each customer’s specific wants, needs, and likely actions. In this respect, the idea isn’t to play God, but to build a better mirror for staring at our own reflections.
Feeding Data into Algorithms
So how will all of this come together?
Data security remains a problem. Consent needs to be better established. Silos limit AI from a comprehensive view and the clearest possible understanding. But there is no existential threat with this industry-specific form of AI. Industry experts contend that, if anything, it will simply mean that marketing is less annoying for the people being targeted, and more cost-effective for companies. Engagement analytics, CRM data, search keywords, shopping cart activity, and other inputs can help mar-tech AI determine intent, separating transactional and informational quests. Then, the AI can serve up quality content that guides each individual along their own buyer’s journey in a meaningful way.
It’s mutually beneficial. A contextually relevant offer is more likely to be clicked on. A well-understood customer is more likely to checkout. Data-driven strategies consistently outperform baseless speculations and scattershot approaches. There is currently a high level of interest in AI, but mar-tech AI is mostly interested in us. And it’s learning from every word and whisper.
