Why Marketers May Or May Not Like OpenAI's Conversational Assistant

I watched the OpenAI announcements earlier this week, and I was floored.  The advancement of Conversational AI and the integration of voice and video features translate to a fully active AI assistant that mirrors what we have seen in movies, with broad implications for the advertising industry.

Advertising is about two things: creativity and analytics.  Both factors are best served when you have an exchange of information between two or more people, because you can build on ideas to create something more resonant and effective.

The advancements I saw this week enable this exchange of information to happen with one person and their AI assistant.  I still maintain that AI is not replacing anyone, but it does make it easier for one person to get into a groove in the same way as two people sitting around a room could do.

Two use cases come to mind for marketers.  First, the creative angle.  I love to brainstorm with others about messages and storytelling.  Sometimes a great brainstorm consists of someone who thinks outside of the box and the other who can build on those ideas, or keep things grounded in a rational sense.  The use case of an AI assistant means you can go back and forth, talking rather than typing, and coming up with ideas off the cuff while the AI plays the “straight” position of keeping things grounded, while providing literary references and concepts that build from your point of view.



In this use, the AI is not being creative, but it can provide answers to questions and references that fuel your creative spark.  The old idiom that two heads are better than one is still quite accurate, but one head with immediate access to an unfathomable amount of information and reference points can be almost as valuable if used correctly.

The second use case comes in the form of analytics.  The visual elements I saw yesterday led me to think of ways the AI assistant can build visuals and create hypotheses based on those visuals.  You can show it data and ask it to crunch the data.  You can show it charts and ask it to provide insights based on what it sees.  While it may not understand the more nuanced elements of media mix, flighting and the interplay of messages from one medium to the next, it can provide ideas and places for an analytical person to begin thinking, guiding them down the path toward hypotheses worth proving.

What’s more, if you use an AI often for your campaign analysis and are consistent with your brand, it just might gain enough insight to be particularly effective for you and your use case.  After all, training data is what makes the AI more efficient over time.

One big question was whether a brand would ever create a more conversational chatbot for itself, allowing consumers to literally engage with its brand.  I think this idea is actually going to scare some marketers away from AI even further.  Marketers like to control their brand, and giving your brand a voice and allowing it to engage on its own is simply too much loss of control.  It opens up a brand to be tricked into saying something that it shouldn’t say, getting it into hot water. 

AI tools can be incredibly useful, but there are just some things that marketers really won’t and can’t allow.  That loss of control is one of them.  The use cases I mentioned above allow the AI to function in a controlled environment where marketers can choose to use or disregard what comes of the conversation -- a process that works far better.

Next story loading loading..