Defining AI Use Cases: Clearing the Confusion for Marketers

Geoff Livingston
5 min readJul 16, 2023


Our most recent No BrAINer podcast picked up this topic.

Let’s dive into the realm of AI use cases, especially from a marketer’s perspective. Use cases are business- and data science team-generated business cases for their AI models. Unfortunately, many marketing AI vendors and parroting pundits and media sell features as use cases.

While I attempted to explain use cases to the marketing community in the recent past, a simpler definition is needed. Use cases are purposeful applications of AI models paired with business cases, including specific and measurable outcomes. Understanding what use cases are, how they’re being discussed, and how to craft them effectively for success is important.

Breaking Down Marketers’ Generative AI Confusion

Some enterprises see generative images as a means to create internal and online content that is too expensive to hire original works and/or license en masse. This can save companies tens of thousand of dollars

There is great confusion in the marketing community between AI, generative AI news, and AI tools such as ChatGPT, Stable Diffusion, and LLMs. This often leads to perplexing conversations about the ideal use cases for individual marketing organizations.

The goodness is that much of the senior marketing leadership community is discussing how to use generative AI. The bad news is that the conversation often begins with writing a blog with ChatGPT or not understanding how to use Discord to interface with Midjourney.

To clarify, trying out ChatGPT or Midjourney or trying Momento to splice video for social media is not a use case. These are great tools, I use them myself, but a use case goes beyond this.

Use cases help organizations plan AI projects with specific outcomes and resources in mind. They provide a measurable framework for success, including ROI, cost savings, work efficiencies, etc., and outline the necessary steps to success.

Unfortunately, many marketers are misled by the promotional language coming from vendors. The way vendors market their AI applications, proposing them as “use cases” or “uses,” essentially confuses use cases with mere marketing actions.

For instance, when proposes potential uses such as social media posts, ad copy, or blog posts as actual use cases. At the same time, while I appreciate what is trying to do, equating their featured prompts and templates as use cases may be overreaching. This confusion is further reinforced when analysts, journalists, and thought leaders (ahem, bloggers) parrot this messaging.

Getting Down to Business

An AI-enhanced human at work? Created on Midjourney.

So, the real question is, how should marketers navigate this hype and confusion? Start by understanding that use cases and prosumer AI applications as not necessarily the same thing.

Think of a use case as an “AI model that resolves your business problem.” While vendors may suggest problems their application can fix, these are not always strategic nor an ideal first step into AI for marketing organizations.

Instead, a marketing leader will want to consider their challenges and where AI-driven automation of tasks may help. Their choice of AI implementations may or may not be generative. Some will see far more impact deploying a machine learning implementation to add clarity and mine its customer journey analytics.

If it is a generative AI implementation, the use case will likely be more complex, with a systematic process and measurable outcomes included in the process.

For example, instead of writing a blog, the AI writer and image creator may instead have a use case where they create 40% more content. That would include creating emails, blogs, infographics, and videos for each campaign while maintaining message integrity to generate 30% more MQLs (marketing qualified leads). Now, that’s more interesting to a CMO.

The Historical Context of Use Cases

AI has been around for some time now, decades in comparison to months for ChatGPT.

Before the generative boom, AI was the domain of data science teams in large enterprises, and use cases were built based on ROI potential and resourcing. Building a Machine Learning or NLP, or multi-algorithm model could cost thousands to millions of dollars, depending on the complexity of the enterprise and its data sources.

However, with the influx of niche generative AI unicorn apps, all departments are scrambling, and individuals can license individual access to prosumer apps. Of course, with the boom have come many dynamic stories of instances where the tools have created instant successes. And these are very exciting!

Please remember your use case is not necessarily a vendor’s use case. Think through your implementation, starting with the problem you want to solve; the business result you want to achieve, the required resources, and whether a commercial off-the-shelf AI or a customized solution suits your needs best.

Some Public Marketing Use Cases to Consider

User generated art featuring Coca-Cola.

Fortunately, AI has been around far longer than ChatGPT. A quick search reveals several use cases. Usually, these use cases show a specific application of an AI model with a tangible, positive enterprise-specific outcome, preferably quantified. Here are a few examples.

Cyber Inc leveraged Synthesia AI, an AI video creation platform, to generate videos for their online courses. This allowed them to quickly scale up their content creation and broaden their global reach by creating videos in multiple languages.

In another case, Volkswagen decided to rely on data to guide their ad buying decisions. They leveraged AI for better forecasting, thereby reducing hidden costs and increasing their sales by 20%.

Coca-Cola partnered with a data analytics firm to develop an AI system to analyze sales data and identify patterns in customer preferences. This resulted in improved sales and distribution efficiencies, more targeted campaigns, and a virtual assistant to aid customers. Coca-Cola also recently initiated a user-generated AI and digital art campaign.

Then there’s online retailer Dixons Carphone, who turned to AI-based tool Phrasee to enhance the success of its email marketing campaigns. Phrasee blends NLP with algorithms to understand customer language and suggests what their email subject lines should say. The result? A 10% increase in open rates and a 25% boost in click-through rates.

As you can see, successful AI is so much more than just the applications and vendors. It delivers meaningful outcomes to enterprises within a defined business use case.

In conclusion, think of a use case as your AI project strategy, complete with KPIs and costs. Don’t half-ass it; this is your chance to establish AI leadership. Start by building your company’s best use case.

As noted in the opening, much of this content was covered in the most recent No Brainer Podcast: ChatGPT Is Not a Use Case. Check it out.