AI-Powered Marketing for Videos

A Recap from National Association of Broadcasters show
Author: Michael Richter
 
I attended the 2023 National Association of Broadcasters in Las Vegas from April 15-19th and it was an eye-opening experience.
 
Generative AI (see video!) was a major theme and I attended many panel sessions and exhibits to engage with early adopters of Generative AI tools. I find that in the media and entertainment industry, and with broadcasters, there tends to be a lot of early adoption of new technology. There was a lot of energy and excitement at the conference and a feeling that something new is afoot. It reminded me of the days I worked for News Corporation, MGM and Turner during the pioneering days of cable and satellite TV in Asia Pacific during the 90’s. Back then, we were pushing the envelope with new technology and expanding into new markets. It was a mix of technology blended with access to new content to the far reaches of the world. The energy today feels similar. Creative professionals are now working with new and more accessible technologies with amazing results.
 
Traditionally, it was just professionals involved with emerging technologies that were in the know. Product marketers would help to communicate and demystify these complex technologies through events, eBooks, blog articles, and videos. I worked as a product strategist and product marketer with blockchain, IoT, AI and experienced how these have been applied to SaaS applications for CX, SCM and ERP.
 
Enter ChatGPT. It’s a whole new ballgame.
 
I believe most of my technology colleagues will agree that the rapid rate of adoption for ChatGPT is unprecedented. I decided to attend NAB in Las Vegas with a mission to figure out how to apply AI to everyday marketing tasks associated with a typical video asset. What I learned was phenomenal.
 
The first step in any project is to know your customer. That requires research. Not everyone takes this first step. Have you ever been in a meeting with a customer and some of the team members in the room don’t know the industry, location, size or current strategy for a customer that you are discussing their next video project with? In the fast-paced world we live in, sometimes there is simply not enough time to get up to speed with even the basics.
 
But now, instead of sifting through web pages, Linkedin profiles or other searches on browsers, now you can create a prompt in ChatGPT to quickly learn about your customer with a response that resembles human language. And sure, there are instances where the result is not accurate and that’s why humans need to be involved, but often you will get something to start with from which to iterate and evolve until you feel comfortable with the result. The more detailed your prompt, the better your result. Why not build this into your pre-production workflow?
 
For videos, story concepts need to be conceived, storyboards created, and scripts written. It can take many hours, days and sometimes weeks to land on a good idea before presenting a draft to client stakeholders. Imagine if you could reduce that time to a fraction of what it was. Multiple iterations of creative concepts can be developed with the help of AI. It’s a way of brainstorming, cross pollinating ideas, and branching off into different realms. Once a concept is agreed upon by the creative team, they can create a draft of a script with a specific number of words such as 150 words for a 1-minute video or 30 characters or less for a title. Need a description for a set of thumbnails? No problem, you can generate a variety of them to choose from by deriving them from the script.
 
From that story and script, marketers can also generate multiple quotes and then ask which one will perform best for SEO. Need a 15-second, 30-second or 1-minute version of a 2-minute video, no problem, that can easily be accomplished in seconds.
 
And then there’s metadata. It’s the information that is used to search and find assets. How can ChatGPT assist? How about just asking it to provide 10 keywords based on the script that you have vetted? Presto! You have 10 keywords to start with versus manually creating them.
 
Sure, it may not be perfect, but in the early stages of drafts and when there are many stakeholders, why not nail down the length and deliver something quickly so everyone can weigh in early and then you will know what resonates before taking the next step?
 
The bottom line is, with the ability to do things in minutes versus hours and days, there’s more time to focus on creative versus the minutia, more time to make things perfect and finesse the concept, and faster time to manage all the requirements to publish and distribute videos.
 
And that’s only scratching the surface because that’s just text-based prompts with text responses. There are many tools being used to work with visual content, facial expressions, voice and virtual imagery.
 
No doubt, there are ethical considerations and concerns. Legal teams at companies will need to determine how to put guardrails on the use of AI and where and how to build them into their workflows. Users need to become more aware of copyright laws and use of likenesses because surely there will be cases of abuse. It’s probably better to play it safe and be conservative rather than to take risks at this juncture.
 
Most major media companies are developing new content using AI but many haven’t taken these public yet as they are working through some of the legal, copyright and IP issues. But the point is, it’s happening, and at a rapid pace.
 
Just for fun, after I wrote this article, I put it into ChatGPT. I asked it to tell me how many words it was, to provide me with three quotes that came out of the article, a title under 30 characters, keywords, a 150 script and even an outline for a video. It took less than a minute to do all this. Do you think it’s time to learn more about prompts for AI?