There’s a lot of talk, but what is actually happening in the real world? What are those marketers & their agencies who are using AI actually doing?
We won’t pretend to know everything that everyone is doing, but here are a few things we’re seeing and hearing about:
LLMs to help sift through the abundance of data
Most marketing teams are operating in a data-rich world. There’ more data available to marketing teams & agencies now than ever before. This is becoming very expense to analyze and make sense of. Some teams are employing LLM’s [large language models] like Claude and ChatGPT to do preliminary organization of data sets. These models are still not reliable enough to do the full end-to-end job of an analyst, but they are quite good at reducing the up-front workload of analysts, allowing them to focus their expertise and experience on much more narrow scopes of data. As time is shifting from “sort this data” to “so what”, it’s helping teams get sharper with insights and driving change in marketing.
Personalization
This is still early days, but there are some prototypes and experiments with using AI to upskill traditional recommendation engines (think product search results in e-commerce sites). Early prototypes and experiments use AI to upskill traditional recommendation engines, moving from “people who bought this also bought…” to more contextual, behavior-aware recommendations. Combining browsing patterns, time of day, and recent support ticket sentiment tailors what customers see next. Some retailers experiment with AI that rewrites product descriptions on the fly, adapting tone and emphasis based on user segments. The goal is genuinely helpful experiences, not pushy ones. Achieving this requires a thoughtful balance of automation, brand voice, and human review.
Leveling up existing models
You’ve got existing models that have served you well for years; maybe it’s some form of econometric model, or a churn prediction and retention model. There are always some decision-making steps in reviewing model outputs and, more common [and less common-knowledge] the refinement steps within model updates & iteration. AI tools are being integrated into these steps to help speed up decision-making and refinement analysis without jeopardizing the reliability of the models. We ourselves have helped a few clients on these kinds of projects, and while nuanced, we’ve seen some amazing outcomes as a result.
Content & creative drafts
If you’ve spent more than 5 seconds on LinkedIn you’ve seen someone post about using AI tools for copy writing and image generation. The results are varied and those that have successfully integrated AI as part of their production process are those who are dedicating resources to making it happen. It’s not really a cost savings exercise (yet) – but speed to market can be as high as 10x (especially for smaller teams with lower budgeted briefs). There’s definitely a diminishing return on this for bigger-budget items, and we haven’t heard of anyone successfully replacing entire creative teams with AI. What we have heard is creative teams who embrace it are working it into their workflows and creative processes to speed things up.
You’ve heard us say it before but for consistency’s sake we’ll say it again: AI won’t replace people’s jobs entirely. Today the impacts that are being actualized are all around augmenting and enhancing existing people and processes. We think we’re pretty savvy to this stuff, and are always happy to [have a chat] (https://onesandheroes.com/contact-us) if you’ve got questions.
