#KatanaTalks on Twitter Live: Artificial Intelligence

#KatanaTalks is a monthly Twitter Live series in which we sit down with digital marketing professionals and ask them how to predict and leverage industry changes to orient marketers and brands for success.

On this month’s installment of #KatanaTalks on Twitter Live, we discussed Artificial Intelligence— how it differs from Machine Learning and how these technologies will affect media buying and ad targeting. We sat down with two experts from A.I. Thinktank Analytics Ventures, Dimitry Fisher and Nicolas Colinas, each of whom brought extensive experience in A.I. research to our conversation about A.I. If you missed the live stream, watch it here! We also have a full briefing on the takeaways below— enjoy!

Dimitry Fisher, Chief A.I. officer of AV Labs

  • What is A.I. and how is it different from Machine Learning?
      • A.I. is the capability of artificial entities or agents to perform meaningful inference from the data inputs they receive from their environment. These meaningful inferences could be in the form of predictions, actions, detection, or classification.
      • Machine learning is essentially a subset of A.I., it is the capabilities of machines, like computers, to make a meaningful inference.
      • Remember: Not all A.I. is machine learning. There are A.I. systems such as web browsers or database systems that are machines, but they are not learning.
      • In terms of the marketing and advertising world—100% of what marketers deal with will be machine learning.
  • A.I. is a broad concept that has been around since the 60’s, why the sudden boom today?
      • Two simple reasons: The volumes of data have increased substantially, and the computing power we have today has risen exponentially since the early 2000’s. 
  • Is there a truth to the belief that AI could potentially make human buyers or creative individuals obsolete?
    • Not really. The A.I. and machine learning tools in the advertising industry won’t dispense human labor, but rather alter the job duties a bit. The mundane work of setting up ad groups and setting bid prices will be replaced, but the creative aspect of this job requires a human. Marketers can now focus most of their attention on the creativity of their ads, and test for which ones will perform better. A.I. will be more of a partner or employee rather than a replacement for marketing specialists.

Nicolas Colinas, Data scientist at AV labs

  • Where does machine learning come into play when it comes to media buying?
    • The ideal case is that A.I. will have a history of every action taken online. Based on this information, it can make predictions of what products a particular user would like. For a media buyer, you can advertise only to a person who is truly interested in your product or service. Basically, you won’t be spamming anyone.
  • What about ad targeting?
    • For targeting, A.I. can help you discover certain segments that you may not have known you had interest from. It can parse out all of the data from all the ads you sent, and see what influences make users interact with your ad. You can also see what made them click and then you can tailor your ad to make it more intriguing to a subset of users. Users are no longer dropped in simple buckets but are more nuanced with A.I. tools.

Katana’s Take: 

We are pretty much in the first wave of A.I and machine learning. There is a landmine of data points out there. For agencies to stay ahead of the game during the  A.I. revolution they need to understand three things:

  1. In order for A.I. to be your employee, your partner, your ally, there needs to be a sense of trust. In today’s media landscape, media buyers currently do not have a level of trust. This is due to not being well acquainted with the technology. Become more familiar with these tools!
  2. You’re are dealing with a lot of different platforms—whether it’s Google or Facebook. Each of these platforms has their own unique customized plan to how they approach machine learning. All with different levels of sophistication. Make sure to discern how each platform will be beneficial for your campaigns.
  3. Have an understanding of the Black Box algorithm. You have to almost trust it blindly today, but if we as an industry keep pushing for transparency from this data-processing device, it will improve trust and help marketers make optimal decisions.

Follow us @katana_media and follow our blog on Medium, medium.com/katana-media for updates on our next Twitter Live show, #KatanaTalks.


Also published on Medium.