Do you ever wonder about all the clicks you pay for that aren’t relevant to your offering? Even if you have built a pretty good exclusion list, you could still be wasting money.
It’s not reasonable to expect a human being (or a team of them) to imagine all the irrelevant terms that the rest of the world might be searching.
But this is exactly the kind of narrow problem that is ideal for solving with artificial intelligence according to Hans Kaspersetz, the CEO of Arteric.
In this episode, we talk about:
- How brands can save money on PPC while providing a better experience by applying AI to large data sets of search queries.
- How Google and other companies are outsourcing the “supervised training” of artificial intelligence to all of us
- The stakes for marketers when Siri and Alexa serve a single answer to a verbal query
By analyzing millions of search terms to identify variants that have the same intent, a company can ensure that it delivers the correct content for each of those. The same technology, Latent Semantic Analysis, can be used to discover the irrelevant terms that may seem related but indicate a different intent.
The progress is stunning. What used to take his team 40 to 80 hours worth of labor can now be completed in 10 minutes or 15 minutes.
This has possibilities for training chatbots as well. Instead of hiring lots of chat operators to respond to people, the chat operators supervise the chatbot and the artificial intelligence so their performance at guiding customers to the correct answers improves as well.