Here at Invoca we were fortunate enough to have Jim Sterne moderate a panel at our Call Intelligence Summit in October. He added a lot of great perspective on how voice and AI will impact marketing, so we wanted to continue the conversation here on the blog! For background, Jim is a true digital marketing pioneer. He began speaking, writing and consulting about the possibilities of using the Internet for advertising, marketing, sales, and customer service starting in 1994, a year after the National Center for Supercomputing Applications (NCSA) released the first public version of its graphical World Wide Web browser. He’s written twelve books, presented at conferences around the world, and lectured at Stanford, University of California and MIT.
1. AI is poised to impact all facets of the customer experience in 2018. What is the opportunity for marketers in particular? How should they be looking at AI?
First off, AI capabilities will be baked into everything we do technically. Five years from now it won’t be called artificial intelligence, it will just be called computing. Our expectation will be that a machine will learn, and every interaction will be better than the last one.
When it comes to marketing, we’ve been doing classic, rules-based automation for a long time. We’ve been using platforms like Salesforce Marketing Cloud or Marketo to trigger certain messages and ads to a targeted set of people. We based these messages and targets on things like demographic data and whether they’re exhibiting certain behaviors, like abandoning a shopping cart. Applying AI to this process means that all of this optimization gets done in real-time. It’s constantly analyzing and improving the campaign results and using that to recreate the model to send the next set of messages or ads out.
For marketers, this could have wide ranging implications – from being able to optimize marketing spend (and not burning money on things that aren’t working) to automatically improving the customer experience by personalizing interactions across devices and channels, online and offline.
2. How are marketing organizations effectively using AI to accomplish their business goals?
On the sophisticated end, they’re hiring data scientists to write algorithms, and building systems and large data warehouses against which those systems learn. However, the vast majority of us are using the tools that are already out there, instead of building this expertise in-house.
For example, if you use Salesforce, you can click the Einstein button to apply machine learning against all the sales and CRM data your company has. Adobe Sensei and IBM Watson are other similar large-scale offerings. However, I think a lot of marketers are successfully working with some of the smaller companies that offer marketing-specific AI tools (several of which won the CB Insights AI 100 award).
If you do this, you don’t need to get deeply involved in the details of how the tools work, and you don’t need a data scientist on staff. Albert.AI is one example of a startup that has demonstrated some immediate and concrete results for its customer, Harley Davidson. Check out the case study here.
3. Voice interactions have come to the forefront over the past year or so. How should marketers be thinking about consumers starting to talk more vs. type?
Between Siri, Cortana, Google and Alexa, we’re getting trained to talk to our smartphones and appliances. We’ve also got chatbots on websites and I think people will soon expect this interaction to be voice rather than typing. People should be able to go to a website and click the “Do you have a question?” icon, ask “When’s the next flight to San Francisco?” and receive an answer in real-time via voice. Bots will also reach a point where if it doesn’t know the answer to your question, you’ll be connected to a human. And the expectation will be that it’s absolutely seamless and that your entire conversation will be handed over to a human before you’re transferred, so that you don’t have to repeat yourself. In fact, recent research by Invoca found that 76% of voice assistant users said that if the device could easily connect them to a human who could answer their question, they would choose this option.
In terms of when voice interactions will become mainstream, several years ago 15% of Google queries were voice and last year that went up to 20%. It’s going to grow by leaps and bounds. We already know that the younger generation is more apt to use voice, and as voice technology evolves and works even better, we’ll see older generations follow suit.
4. How should marketers think about voice in the context of considered purchases?
Trust plays a huge role. When you go to two different websites that are competing for your considered purchase decision, and you’re dealing with clicking buttons and having conversations with a robot, the thing that’s missing is the impression of confidence. This is much easier to gain in-person or over the phone, which is why people still pick up the phone. When we’re talking about AI and how it applies to considered purchases, there has to be a whole lot psychology and context built into these systems so that you feel comfortable talking to it and prefer it, rather than demanding to speak to a human.
It will take some time for this level of AI to become mainstream, but I think the ‘switch’ will be very clear when it does happen. For example, in 1993 I was on the fence about using the Yellow Pages or this new thing called the World Wide Web. Even though United Airlines had a website, the customer experience was terrible, so I still called to book a flight. But it finally got to a point where it became easier and faster to do it online that over the phone. This will happen with bots — there will come a time when it becomes easier to do most everything (not just checking your bank balance or the weather) via a bot.
5. Do you have any tactical advice for marketers looking to get their heads quickly around AI?
First, go play with systems that are out there. Most are still pretty new, so don’t have high expectations. For example, Levis Strauss created a virtual stylist feature on its website for clothing recommendations. Is it AI? Maybe. But what matters most is the customer experience. Go play with bots and any technology that’s looking to streamline the customer experience and form an opinion on what you love and what you hate.
Second, read up and learn. You don’t need to find a Stanford level course on AI. As a marketing professional, you only need to understand the underlying concepts so that when it comes time to implement a tool at your company, you can speak the language and have an understanding for how it will impact your business. It’s also critical to understand how machine learning differs from automation.
In addition to my new book, here’s a couple blog posts I’ve written that provide a good intro to AI for marketers: