Questions to Ask Before Using Gen AI

Dec 12, 2023

A marketers job is to create stories for the brand; however, with the increase in volume of content required out there to be on so many different channels, it’s so hard for the modern day marketer to keep up. Especially with budgets that are not growing to keep pace, marketers are faced with the issue of new constraints based on traditional ways of sourcing content.

Some of the common challenges that we hear marketers face are: 

  1. Volume & variety: “Getting enough volume and variety of creative is super expensive so we are often limited to a handful of images we need to re-use over and over again”

  2. Localization: We struggle to create localized content because the stuff from our global office does not reflect our local demographic and cultural nuance

  3. Budget: I don’t have budget for an agency that can produce images, so I have to default to stock photography which means my creative is generic by nature and constrained 

This guide is to help digital marketers assess their “readiness” for Gen AI and look at fundamental questions before integrating it into their workflow. These questions help ensure that when you use services like AeroBoard, it is a smooth working engagement. 

We’ve broken down this document into Category of Questions, the Questions themselves, and some Comments from AeroBoard’s POV based on our experience working with other pharma, healthcare, and CPG brands. 

Legal

  • Any copyright or IP concerns?

  • Any restrictions on usage of the images created?

  • Do we need to mention to the public that they are generated images?

The ‘models’ are not real people and the ‘environments’ are not real places. No releases are required. AeroBoard offers an unlimited, global, exclusive license across any channel. Practically, no restriction. Mentions of generated imagery is up to you and your legal team.

Brand

  • How do AI-generated images align to our brand? Will they be consistent? 

  • How do we ensure the quality of the images? Are there QA checks? 

  • What are some limitations of using Gen AI for image creation?

A top tier Gen AI service will be able to leverage your existing brand guidelines with technical detail to execute new productions. For QA, it is no different from normal workflow as there may be people checking at certain levels. There are definitely some limitations such as getting too hyper specific on a scenario. We often try to mitigate that by making sure briefs have some wiggle room for creativity as long as they fit the master theme and brand guidelines.

Localization

  • How might we work with Gen AI to produce more localized creative that is more relevant to consumers? 

  • How should brands work with Gen AI services?

  • Should we allow one region to use it before the rest of world?

Localization is a problem that has existed for the longest time. It still exists. Gen AI just allows you to re-create hero assets from global at a local level (i.e. people that look local, backgrounds, artifacts, etc). This is a new superpower that brands should be aware and take advantage of. Generally, when new technologies for creative production have arisen in the past, they are best contracted at the Regional / Country level, but implementation happens at the Brand level. This ensures enough hands on the relevant creative while not overcomplicating contracts / procurement. For regional testing, it is no different from any new technology or service. Pilot it in one place and then scale. Likely pick a market that has a clear need and criteria for success as starting point.

Value

  • How much is it costing us to license stock images (usually not in perpetuity)? How much does it cost to shoot custom full production? How does that compare to using Gen AI?

  • How fast will we see value?

Do the price/value comparison. In our case, we want to deliver custom, bespoke content at the price of stock licenses. This gives you the ultimate partner in creating content as you scale a brand. If you only need a few assets here and there, we are not likely the best partner; however, if you require a breadth of imagery across multiple brands, that is where we shine in process, people, and technical capabilities. We historically have saved pharma brands 100s of thousands of dollars on production shoots and licensing on images. Within 1 week you will start to see images produced. This would take weeks to months on traditional ways. This allows you to test and learn at a way higher velocity, especially on any programmatic ad buying.

Clear No Go’s for using Gen AI: 

  1. Hyper specific product and use case (i.e. complex surgery device in an operating room)

  2. Only needing a small handful of images for one time use

  3. No urgency on turnaround time and no need to reduce costs (to get more throughput)

  4. Industry where Generative AI is completely vetoed