The age of AI-powered search has human at the center
Artificial intelligence is set to become a disruptive force in the search industry. No surprises there then. As competition for consumer attention escalates, advertisers and publishers are in a race to find the right signals to best understand who a consumer is, what their behavior is, and what their intent will be. Businesses that are responding to this change are starting to use AI and machine learning to uncover new intent-based insights and nurture a new level of personalization, according to a panel discussion organized by The Drum, in association with Bing, in Austin at The Drum Arms.
The panel included Microsoft Advertising’s head of brand studio, Geoff Colon; Managing director, search and social media at Catalyst, a GroupM and WPP Agency, Jim Kensicki; and RGA senior connections planner, Christine Song. It was moderated by The Drum’s editor, Stephen Lepitak.
Unlocking the power of AI to transform search
Sometimes, a simple keyword search doesn’t quite quench our never-ending thirst for knowledge. And that is why, says Colon, search is attempting to find new ways of responding to queries and offering new ways of exploring information. But are brands prepared to capitalize on this shift?
For Song, the pivot will come when we start to see how AI in search is going to “turn marketers and those who specialize in SEN and SEO into more strategic roles rather than figuring out key words.” Intelligent search will inevitably transform the landscape of search, according to Song, who adds: “You need to know what the consumer journey is like, it's less about the exact key words and more about who these people are, what they like, what are they looking for and using that knowledge of them to be able to target them better.
“There is also the fact that AI might make AB testing almost an outdated way of doing creatives. With machine learning we will be able to test several variables and different creatives all at once, instead of having to run multiple experiments just testing one variable.”
For Kensicki, AI can process and interpret patterns in data at a scale people could simply never replicate, making it an essential complement to any search strategist. He says his agency clients are using it “to bring the large data sets and marry the in-market audiences that we have through Microsoft and through Google with offline data. Trying to understand within each of our clients’ customer segments how their online behavior differs and how to use that data to then drive decisions, like where you display media plays or traditional media is a key area in which AI can improve SEO performance.”
Search: the next frontier in creative ideas
In an era where marketers are moving from the age of information to the age of experience, search marketing has been trying to provide brands the opportunity to determine actionable outcomes for consumers. “AI, then, helps us to develop new ad products at a faster rate, moving beyond ‘hey here's search extension ads’ that's antiquated at this point because you want to use AI to figure out what the new solutions that you can involve. How do you use image? How do you involve sounds? How do you involve moving visuals? AI can power those solutions quicker and you can test those and market faster so you can figure out if advertisers want those and then release those to markets. It's speeding up solutions to the market place. It’ll be interesting over the next couple of years to see what the big players do because they have large data sets and the ability to test a lot of these. What will Amazon, Google, Facebook and Microsoft release out in the ecosystem that might look new but also be a remix of things we remember from the past?”, says Colon.
Song says that the “excitement” in our industry will start happening when we realize the potential for AI and search becoming creative. Search is everywhere, it's Google, Bing but also social media; Facebook, Instagram. We search using Google Home, Alexa, Siri, it all will become the next frontier in creative ideas.”
She points to the example of the search campaign featuring Trevor Noah from The Daily Show. When the show was first being launched, the idea was to increase his awareness. A search campaign was created that revolved around the curiosity about Trevor and rewarded those interested in learning more about him with content. To execute this campaign, the campaign leveraged the normal user behavior when seeking to learn more information about someone: searching it.
Search and human behavior
And that is why understanding user behavior provides the many cues required to improve the relevance of search results, the panel agreed.
Colon adds: “One of the things that is most fascinating and fun at SXSW is looking at human behavior. For instance, all the people staring down at their phones and still using their thumbs to find things. That will become antiquated with AI. Your camera, for instance is the key to unlock the world around you ultimately. Pinterest is always considered a social network, but really it is visual search and there's not many companies utilizing that.
“Ultimately search should not be front and center in your life, it should be immersive, it should be wallpaper, so that you can be more human and do things and become more productive.”
Both Kensicki and Colon say that voice search is another such user interface, and even though voice and visual search may not have fully evolved just yet, it is only a matter of time before they become integrated into the consumer’s decision journey.
“AR and machine learning, computer vision is increasing rapidly so fast these days. It's going to be a short amount of time before you can literally use your camera on your phone to search for the things that you want,” according to Song.
The panel agreed that AI plays a critical role in transforming businesses around the world by enabling more informed, personalized and smarter user experiences. An audience however questioned whether AI could solve the consequence of personalization that is the ‘filter bubble.’
When users don’t get to see the information that disagrees with their viewpoint and thus isolating them in their ideological information bubbles, is a major problem agrees Kensicki. “We saw what happened with Facebook over the [Trump] elections, there's the same possibilities for that to happen in search - it’s being driven by advertisers and media dollars - much like social media is. We have the benefit of having seen it happen and learn from it. What Bing has done is being much more thoughtful about the market audience that they use and how they introduce audiences to advertisers in a way which is much more thoughtful. We just need to be more aware.”
Colon admits that Bing has had its share of problems, but it has been working on it for some time. “In some respects, it almost puts tech in a journalist’s role, which is an ethical discussion that we all need to have.”
The future of AI and intelligent search
Meanwhile by 2020, there will be 20 billion devices connected to the internet. At the same time, half of searches will be made by image or voice, while 85% of customer relationships with an organization will be interacted without a human (source: Gartner). So there’s obviously a huge responsibility for the industry to make sure it provides trustworthy and meaningful results.
Colon agrees that with machine learning comes a responsibility to ensure the recommendations made by any search engine are credible and ethical. But perhaps one of the trickiest and perhaps most essential safeguards is protecting that AI from bias.
“When it comes to bias, do you have a diverse team to work on the algorithms that create search results? At Bing, we are asking ourselves that. Sometimes there is a tendency for people to think that algorithms are not biased, and that's just not true. It's based on the creator,” says Colon. He adds that there’s a big push from Microsoft chief executive Satya Nadella, in terms of how to get more people from around the world to work on making search more inclusive.
Kensicki concludes: “We’re lazy and impatient as a population. So the quicker we can serve the right answer, the best piece of content without having to click, it will only help that platform.”