What The History Of Search Tells Us About The Future Of AI
This article was first published on Contagious.
In the wake of AlphaGo’s recent victory over top Go player Lee Sedol, we’ve seen a lot of chatter about artificial intelligence, better known as AI. For consumer, and especially non-Go players, that typically means voice-activated bots, like Amazon’s Alexa or Facebook M. For brands, that brings up the inevitable question: if our customers will be talking to their refrigerators, what will they say and how should we respond?
Whenever you have a new technology like this and you want to know where it’s going, it’s a good idea to look for a close analogy in the past. In this case, we have a pretty good one: search. To see why, think for a moment of how you search. You might think that if you have a question, you simply enter it into Google and get an answer.
But that’s not really what happens. If you want to make sangria, you probably don’t type, ‘How do I make sangria?’ into a search engine. More likely, you enter ‘sangria recipe’. Then if someone spills a cup of it on your carpet, you don’t search on ‘how to clean up sangria’. Much more likely you enter ‘how to remove wine stains’.
These are learned behaviours. Over time, you have learned that if you want to make dinner, the search engine is going to give you more and better options for ‘recipe’ than a natural language query. Likewise, you know that you’re more likely to find exact information on cleaning something if you put ‘how’ in your search. And you also know that you’ll get a more useful result if you enter ‘wine’ instead of ‘sangria’. That’s because over time you have developed an advanced understanding of how Google works.
In other words, not only has Google learned about how people search, you’ve also gotten much better at getting the answers you want. Voice-based AI will almost certainly develop in a similar way. It will not only require the machines to get smarter about understanding natural language, but also force consumers to do the same about AI. The only real difference is that voice assistants will be able to do much more than search engines.
For brands, this is not a minor consideration. If people will learn and grow with these technologies, we need to get into the slipstream of the development now, so that we’re not left behind. In doing so, at least the following four things will be critically important:
Adding ‘do’ to your vocabulary. One big difference between search and voice is that search has predominantly been about asking questions. Voice certainly will have the capability to do that too (at least 41% of adults already use it). But we will also see a new behaviour around getting stuff done. We will learn how to order the assistants to turn on lights, buy paper towels, watch Monday Night Football, and wake us up at 5:30 a.m. As a result, brands will have to get used to being ordered around and include that capability in their products.
Stressing value, not message. Brands are very good at delivering message. However, voice assistants can also provide value to your audience. That’s why many are focusing ‘skills’ on Alexa, which are essentially built-in apps that allow people to do things more easily. A good example is that you can now use voice to quickly order Dominos or an Uber on Alexa. Finding ways to add value like that will be critical to brands’ success.
Integrating existing systems (or finding a partner). Right now, voice might seem like highly advanced technology, only within reach of the Googles and Facebooks of the world. That’s simply not true. Plenty of tools, like Dynaspeak and CMUSphinx can add voice capabilities and basic AI to your existing apps with a fairly reasonable amount of effort.
Above all, learning. Many CMOs are with good reason reluctant to play with a technology for its own sake. While I’m not saying voice will contribute immediately to your bottom line, it is something where you’ll want to have a solid foundation in the future. It also doesn’t require massive investment. A high school kid, for example, recently created a voice-activated chess game by bolting together existing technologies (see below). Simply hooking up your existing properties to voice could yield data on what works and what doesn’t for your audience – and lay the groundwork for future success.
In other words, voice is coming and bringing with it a learning curve and usage patterns that will change substantially over time. Those who get in early will have the best chance of getting out with a good understanding of how it works.
A final point is that marketers should start using the technology personally as well. Too many of us took too long to get on Facebook and never really caught up. My advice is to run, not walk, and get an Echo or similar device. They’re inexpensive and in addition to being the future, they’re also kinda fun.