AI is on the cusp of mainstream adoption, yet there is still a lot of confusion around what it is and how we can use it in our practice. So, Content Magazine sat down with one of North America’s leading experts on the topic Susan Etlinger to find out more.
Etlinger, Altimeter Group’s industry analyst who advises global clients on how to incorporate intelligent technologies into organizational culture and practice, shared her insights on why AI is coming to the fore now, ways marketers can utilize it, and how to get ready for voice search.
Content: What is AI? Why is it becoming so important now?
Susan Etlinger: The first thing to understand about defining AI is that nobody agrees on what it is.
Technologists don’t agree. Business people don’t agree. But I’ll give you a couple of definitions I’ve heard that I think are really useful.
In some ways, people’s view of AI is very aspirational, meaning that a lot of people will look at ‘Black Mirror’ and other films in popular culture, and think of it in terms of sentient machines. That just doesn’t exist, and that may never exist in the way that we’ve formulated it.
AI is not magic. It’s technology that enables machines to sense information, classify it, analyze it, maybe make inferences based upon it. Predict, act, and most importantly learn. The two kinds of characteristics that I think are most important are that it is somewhat autonomous, meaning that it makes decisions on its own. Secondly, that it is learning from data and past experience.
Why is AI coming to fruition now?
There are three reasons.
One is, we have just a tremendous amount of data at our disposal now, digital data that we didn’t have before.
IBM has been saying for years that 90% of the data in the world was created in the last two years alone, and that two-year period may actually be dropping to a matter of months. We are just creating so much data.
Data is the lifeblood of machine learning algorithms, or what we think of as AI.
The second reason is that the cost of computing and processing that data has just dropped so precipitously that we can now process massive amounts of it without having to be a government agency, or a multi-billion-dollar company.
Third, when you add tons of data and the ability to process it, of course that leads to research advancements, and just the improvement of various algorithms that are used to doing different things. It’s a virtuous cycle.
Which types of AI will marketers be able to use?
There are three main areas that I think are relevant for marketers:
1 — Predictive Analytics: Predictive analytics is having the ability to do analytics based on data, and predict future outcomes, and try to figure out, for example, what might be the most efficient way to target an ad? What might be the most efficient way to personalize a campaign? What might be the best type of campaign, or the most engaging content?
2 — Natural Language: ‘Natural language’ can encompass anything from the text in a news story to a customer service chat log, to social media content, what people are posting or blogging, even to voice data; it’s the ability to recognize speech and turn that into meaningful content.
We are beginning to move away from the first and second iteration of the Internet, which consist of URL and app-based content.
And now we are starting to see the capability to use speech and text, chat, and potentially even gestures, to interact with products and services and companies too.
3 — Image Recognition: Computer vision is the ability to understand the contents of an image: a photograph, video, even a drawing or painting.
Image recognition helps us understand context, what’s happening in an image. How many people are there? What’s the weather like? Are they in the country or are they in the city? Are they at the beach or the mountains? Does it look like they are having fun?
That becomes very interesting in terms of content development and personalization.
Gartner predicts that 30% of all browsing will be done without a screen by 2020. What should marketers be doing now to meet the needs of voice search?
Marketers should learn from what’s happening in the customer service world, because when it comes to things like chat and voice, particularly in chat, there is a lot of activity around developing chatbots that can assist people with customer service requests.
There are a couple of reasons for that.
If you are a large enterprise, you’ll have a ton of data on the way that your customers interact with your products.
You have call center calls and chat logs. You have emails. You have lots of data that can be used to better understand the trends and problems that customers have, which you can use to train chatbots to recognize, predict and respond to in a structured way.
A number of organizations start with customer service and then move into marketing. They begin by solving somebody’s problem, find out more about their needs, and essentially deepen that relationship, so they can earn permission to sell, cross-sell, and upsell to them.
How will the advancements in computer vision alter the way brands and businesses promote their products and services?
Computer vision is beginning to be integrated into digital marketing offerings like a marketing cloud, such as Salesforce or Adobe. Or it’s integrated into social media analytics offerings.
In additional to being able to analyze the sort of text data that the customers supply, via Twitter, Facebook, apps or websites, computer vision is actually able to see the images that people are posting about their experience.
This technology could be used to monitor moments of consumption; so, for brands that are very consumer orientated, for example like Coca-Cola, they can see from a photograph that somebody is enjoying their soft drink on a picnic.
Here they can see their brand in the context of other brands. That might provide some input to marketing campaigns, and give them an understanding of what people’s expressions are, and what their temperament might be around their brand.
Can you give an example of AI already in action?
We use AI every single day. If you have a phone, and you are connected to the Internet, you are interacting with AI in one way or another. You are using predictive text on your phone that is trying to anticipate what you are going to say.
If you are using Google or another search engine, or if you’re on Facebook or Twitter, the algorithms determine what you see in your feed.
A lot of positive work is also being done to use computer vision and natural language together to improve accessibility. These things, the predictive cases, the chatbot cases, the voice cases, these are all real and happening now. It’s just that we are in a very embryonic stage of their development.
As AI tech becomes more sophisticated, how too will the practice of marketing?
When I think about the future, what I really hope is that we don’t use machines to replace people. We should get really smart about the things that machines are good at, and really smart about the things that people are good at, and create complimentary relationships.
We shouldn’t automate things for the sake of automating them. We shouldn’t make machines seem human for the sake of making them seem human. That can be off-putting, or even violate customers’ trust.
The real purpose of AI in marketing is to allow for a much more organic series of interactions between people and the organizations that they interact with.