Smart marketers are already making great use of enterprise AI applications and machine learning in marketing. But how?
As quickly as it's being adopted, artificial intelligence still proves challenging for many. The potential for artificial intelligence to revolutionize your business is there, if you can figure out how to tap into it in the way that makes the most sense for your employees, stakeholders, and customers.
Artificial intelligence applications are systems that normally require human intelligence, but are powered by machines.
(If I'm losing you already, this primer on What Is and What Isn't Artificial Intelligence is a great place to start out.)
And where are smart marketers putting artificial intelligence and machine learning to work? In a word: everywhere. As Adam Coates, director of the Baidu Research Silicon Valley AI Lab, said:
"Deep learning is, in some sense, not a product by itself. And so what's going to happen, I think, is that it's going to be in enterprises everywhere, in applications all over the place-whether you're running a data center, trying to help someone with a self-driving car, or trying to forecast the weather or the kinds of crops that are going to grow-all these crazy things. All of those are gradually going to be feeling these ripple effects from AI."
But let's get a little more granular and have a look at how artificial intelligence and machine learning are being used right now by enterprise brands to achieve a number of business and customer satisfaction objectives.
The five areas we're going to dive into typically require human intelligence to navigate, yet innovative brands have put enterprise AI to work in:
Here's how.
Speech recognition sat stagnant for much of the last two decades. It’s only recently that investments in Siri, Alexa and other speech-enabled technologies inspired expanded vocabularies, more natural language, and exponentially higher answer quality. Now, voice recognition powers everything from Google Search to GPS maps and other hands-free systems, to in-home assistants like Google Home and Amazon Echo.
You can make your brand’s content more appealing to voice queries by:
Wherever possible, incorporate speech recognition in the operating model of your own products or applications, as well. Your customers and prospects will appreciate the hands-free convenience.
So, who’s doing it right? Take a look at pizza brand Domino’s, whose stock is up 5,000% since 2008.
The nearly 60-year-old brand has shown that an old dog can indeed learn new tricks, as it’s harnessed the power of new technologies better than perhaps anyone else in its space.
Domino’s first rolled out its Siri-style “voice ordering” system with voice-activated personal assistant Dom in 2014. A couple years later, the pizza chain joined forces with Amazon Alexa to facilitate ordering by voice over the in-home assistant.
Patrón Tequila also partnered with Alexa, for its voice-activated Cocktail Lab recipe library. Patrón also offers a voice-activated Bot-Tender, an AI-enabled chatbot that suggests cocktail ideas to customers on Twitter Direct Message, Facebook Messenger, and the company’s website.
Forrester analyst Collin Colburn recommended at the recent Forrester Consumer Marketing Summit that brands begin their foray into voice by asking two simple questions:
There are applications today that can turn specific information sets and data into content you’d swear was written by a living, breathing human. Companies like BrightEdge are among those leading the charge in machine-created content and automated personalization for the customer journey.
For the most part though, you can expect that AI will aid in content creation, but human marketers are still absolutely necessary. Marketers are needed to create editorial content, choose the right imagery to complete the story, and pair the right images with the right messaging for display ads. The most successful marketers in the near future will embrace AI and its capabilities for cutting down on production time and better informing and personalizing smart content creation, through data insights and analytics.
AI can also be used to personalize entire customer experiences, as in the case of Lufthansa’s AI-enabled features for travelers. You might receive a personalized offer for lounge access between connecting flights, for example, if the algorithm determines that this service would benefit a traveler with your data profile. Repeated rejections “teach” the machine not to offer that service anymore to that individual.
Home Depot’s personalization by location is less granular, but still tailors the customer experience by using a shopper’s location to tie in localized design trends and products. A customer walking into a store in Detroit will receive a different set of recommendations than a shopper browsing the app in Florida. Each store is geofenced so that when you physically enter a store, it recognizes that you’re in shopping mode and begins to display product locations. In some stores, customers can even access guided directions to specific products and aisles.
Recommendation engines like these use AI to “learn” how to better serve customers in their decision-making moments, informed by more data than a human could ever hope to analyze. These technologies go far beyond simply making a recommendation based on some set rules programmed in early on. Rather, today’s recommendation engines draw conclusions based on myriad data points drawn from numerous sources.
Chatbots and intelligent agents are all the rage, and it’s no wonder. Some have become so convincing that you’d have a hard time telling them apart from a human customer service agent (who, after all, is likely following a script and pulling solutions from a canned collection of pre-approved options anyway).
Chinese search giant Baidu launched its voice-activated intelligent agent fairly early on, in 2015. At launch, Duer was introduced as an adorable robot, but the technology behind it was quickly integrated into Baidu Mobile to enable product ordering from within the app. Within a few years, Baidu Duer became a standalone business unit and today, more than 2,000 engineers across the U.S. and China are “working collaboratively to advance AI to the next frontier.”
Chatbots are being integrated into every facet of customer service and communications, including across Facebook Messenger, Slack, Twitter and more. A voice or text-activated chatbot inside the Starbucks app allows you to place an order and completes the transaction with pricing and pick-up time information. You can find info on your recent Mastercard transactions through its chatbot in Messenger, and even call for a Lyft from your Facebook Messenger, Slack or Amazon Echo.
In-home intelligent agents are now a dime a dozen and many consumers that use one prefer to use intelligent agents to discover more generalized information. This presents a great opportunity for brands to “be the answer” for those consumers in a discovery state of mind, by creating authoritative, top quality content.
Predictive analytics take conversion management to the next level by helping to pull potential outcomes from the mass of data you have at your disposal. This can involve analyzing inbound communications and traditional metrics, like consumer engagement, closed business, and communication channels.
Rapidminer, Birst, and Sisense, to name a few, are helping marketers make smarter, data-backed decisions. Consider the unmanageable volume of marketing data the average enterprise brand is contending with, including clicks, views, time-on-page, purchases, email responses, and more. Analyzing this data in a timely way in order to uncover actionable insights is a great feat, indeed.
That’s just one piece of the puzzle, though. As my fellow columnist Mike Kaput wrote, it’s important to understand how prescriptive analytics works, as well. As he says, “A predictive sentiment tool will tell me what sentiment score a bit of text might have based on past data. A prescriptive sentiment tool will tell me what to do to improve that sentiment score.”
This has the potential to push the capabilities of your marketing technology far beyond simple automation, to recommendation and even implementation. Few tools are able to do all three, but Mike shared Phrasee as an example you’ll want to check out.
“Phrasee assesses email subject line performance based on open rates. It recommends new subject lines that will be successful based on this data. And it implements subject lines, automatically writing them.”
AI takes us far beyond human limitations, to the realm of intelligence on a scale no human can possibly muster. Algorithms are learning to think; to not only predict, but to act on those predictions. And each time they do, they become smarter as a result. Again, it can all sound pretty scary for a modern marketing professional. Rather than taking over our jobs though, these technologies have the potential to augment and enhance our performance.
We’ve known programmatic advertising in many forms for some time now—think search engine marketing on channels like Google AdWords, Facebook or Twitter. An entire cottage industry with companies like PredictiveBid and Albert have sprung up to service programmatic buyers and sellers.
As its heart, programmatic is simply the automated process of buying and selling ad inventory through an exchange. Gone are the days of the long sales lunches and manual orders. Today, it’s all about real-time bidding for inventory across mobile, display, video, social, and televisions. Increasingly, bids are based on AI-enabled insights.
Algorithms might analyze a site visitor’s behavior to serve up personalized ad content in the moment. That information is likely being gathered and used to inform future optimizations and ad content creation, as well. Demand and supply-side platforms (DSPs and SSPs) and data management platforms (DMPs) collect all kinds of first and third-party data, to inform decisions on personalization and ad purchasing.
Ad sellers will be able to focus more intensely on outcomes and on helping their clients achieve greater ROI through personalizations driven by ever-increasingly personal data. Through AI, psychographics will pick up where demographics leaves off, to give greater insight into the motives, needs, and purchasing behaviors that drive consumers in their decision-making moments.
Volkswagen found that using AI for its ad buys in Germany resulted in better performance than even its agency could provide. Whenever Volkswagen uses the recommendations from Blackwood Seven, a Danish media agency that uses AI and predictive analytics to forecast ad spend decisions, it sells more cars than it would have if it had gone with its media agency’s recommendation, the head of marketing for Volkswagen’s passenger cars in Germany told Digiday. In some instances, the difference between the algorithm’s and its agency’s car orders has been as high as 20 percent.
Artificial intelligence is advancing beyond data analysis to smart decision-making, as machines learn to more accurately mimic human traits and behaviors. We have access to more data sources than ever before, as even voice and video have become mineable. Enhanced analytics paired with machine capabilities to learn, plan, and even execute will help marketers improve both the efficacy and effectiveness of smarter campaigns.
Look for AI applications that use artificial intelligence to power more meaningful customer interactions, transactions, and experiences. Increasingly, the tools available to marketers will have AI capabilities built in, as the vast majority of marketers expect their marketing tech to be -powered by artificial intelligence in some fashion. When BrightEdge asked if they expected their marketing tech provider to have native AI capabilities, just 9% answered "no."
As for teams, look for hybrid marketers with equal parts agility, technological proficiency, and inherent accountability, who can think and perform strategically across multiple disciplines.
The ideal marketing professional in an environment that prizes artificial intelligence is able to move seamlessly from the creativity of content and social to the science of optimization and analytics across AI applications.