The Shapeless Nature of AI
Attending MAICON this week in Cleveland, I landed on a simple way to explain how AI works that resonated with attending marketers: Like water, AI has no finite shape or use type.
Continuing with this analogy, like water, AI takes the shape of the container you put it in. The container is your AI model as defined by a use case.
Water can take any shape if you can build a container in that form. Enterprise executions of AI are only limited to imagining the strategy, model, and resources needed to make it come alive. Thus a use case puts business logic behind an imagined application.
When water is frozen, it takes the shape of the land or container that holds it. When an AI model is built, it becomes an application that performs tasks defined by the algorithms and governance structures created by a data science team.
With an artistic vision in mind, you can take a chisel to ice and shape it to your creative ideal. When you train an AI model with your data, you teach it to become practical and provide answers and results specific to you.
Remember that data science teams usually include subject matter experts to help shape the model and then test their models with AI-savvy users to support their vision take shape. This type of user-driven knowledge makes for valuable applications that hit the mark.
What About Vapor?
Water can become vapor, absorbed into the air through evaporation. Unfortunately, AI can become vapor, too, mainly when vendors sell their wares as panaceas that perform full-on marketing campaigns, predictive analytics in ABM, etc.
Indeed, there is water in the air. We can see it in the clouds. But it is only tangible once condensation occurs, either as rain, snow, hail, or dew.
Similarly, vaporware only becomes a reality when function meets fiction. I think promised AI is a concept. It needs to be executed to become tangible.
In many cases, we can see the clouds in the form of prosumer AI applications that offer enterprise panaceas of fully executed marketing (or sub in your professions of choice) tasks.
Yet, the AI rain is hit or miss. Getting the correct outcome takes a lot of guidance from prompting from humans to get the AI to produce a meaningful result. Sometimes, depending on the AI, the outcome doesn’t cut the mustard.
Make Your Own Water
Unfortunately, for many in the business community, we are being sold half-baked AI by vendors. It’s essential to take a step back and remember that AI takes the form we want it to within our organizations.
It’s essential to question what vendors are selling… And not just from the standpoint of, “Will their [so-called] use case will work?”
To make AI successful, we must define how AI can best work for our companies. We must express and implement our own use cases. What’s the best impact AI can produce for my marketing organization and business? Is the itch worth the scratch?
Every marketing organization can benefit from AI. Marketing organizations should take the time to find a meaningful fit that will help them perform their tasks better.
That begins by identifying painpoints and needs, from generating leads and revenue to building brands and educating stakeholders. Then building a use case and model to resolve it.
AI is what we make of it. Not what we are sold.
GrammarlyGo was used to proof this article.
