Skip to contentSkip to footer
People
GenAI’s role in retail continues to grow, and is here to stay. While recent tech advancements like crypto and NFTs didn’t live up to the hype, GenAI is already driving real business implications that will soon expand—OpenAI’s recent developer conference brought exciting updates.
Many retailers have already implemented GenAI for content creation and the automation of routine processes, especially around call centers. But we believe the greatest impact will come from mixing GenAI with machine learning (ML) in innovative ways. ML focuses on developing algorithms that analyze complex data and identify patterns, which inform predictions and decisions. Generative AI takes machine learning techniques and uses this input to generate new, original content—like text, audio, and video—from the teachings of its training data.
A hybrid approach can support the level of hyper-personalization enabled by micro-segmentation and sequential recommenders, and create a virtuous cycle that drives continuous improvement. As such, retailers that blend these technologies into existing processes will find their impact multiplied. When implemented within a broader marketing effectiveness program, these capabilities augment customer count, basket size, and total trips—and can drive a meaningful revenue uplift.
For example, we recently worked with a nationwide specialty retailer to develop a suite of custom marketing solutions, informed by ML and GenAI, to improve the efficacy of promotional emails. The solution provided customers with a personalized experience that increased open and click rates by 30-45%; the company tied the campaign’s results to revenue improvements.
There are many applications for this combined approach, so choosing the right parameters and setting up the right process is key. Let’s use an example of a common marketing challenge: how to generate additional revenue with limited resources.
Leverage AI and ML to pinpoint the right target customers
Sales and marketing teams often have a laundry list of priorities with limited touchpoints and resources. Ensuring the right customers are prioritized can increase overall efficiency—AI advances make this much easier.
Use customer lifetime value (CLV)—estimated by ML models that forecast customer churn and future spend patterns based on each customer’s purchase experience—to identify your high-potential customers. Start by considering how you can develop proprietary CLV models that help your marketing teams with prospects. Work to define any constraints that could limit your model’s effectiveness.
ML algorithms can then help you build response propensity models that discern which high-CLV customers are most likely to make additional purchases after contact. Beyond purchase intent, these models will also teach you what these customers are most likely to buy, and which segments might be more promotion-sensitive than others.
To dial in your strategy, test the quality of your model and customer response. One company taking this approach generated 40% more revenue per targeted customer and double the average order value by contacting high-CLV customers provided by its ML models compared to traditional targets.
Utilize GenAI to craft and test messaging
Knowing who to contact is only half the challenge. At this stage, the ML models hand off to GenAI to create and deliver personalized offers and messages that motivate customers. Unlike traditional recommendation systems, GenAI product recommendation engines generate content and suggestions based on each user’s preferences and previous interactions. This takes personalization down to the individual customer level, a key advantage for marketers and sales teams looking to reach customers along their personal journeys.
You can utilize your ML output to develop micro-segments that allow your marketing team to hone messages to specific sub-groups. Then, feed each micro-segment into GenAI algorithms to create different subject lines and messages tailored to each group. Test all iterations to uncover what resonates best.
When implementing this approach, make sure your teams are aligned so when moving from ML models to micro-segments to testing, the key data and “why” behind each step properly transfers. A quality control focus will ensure all your models rely on a shared understanding.
Iterate and improve moving forward
This approach allows you to quickly optimize your most impactful messages. Practical reinforcement learning has long been a business challenge, as humans can only create so many iterations and can’t always discern the “why” behind a message’s success. But with ML, you can utilize the knowledge gained via testing to take the top-performing AI-crafted messages, iterate on those, and keep testing and improving.
This creates a virtuous loop that boosts message effectiveness with time. Your ML models, providing high-value CLV customers, will continue to feed pinpoint target segments into GenAI algorithms which will tailor more and more effective messages to customers—a holy grail for your sales and marketing teams.
Going forward, evaluate how else you can apply GenAI and ML to your processes. Consider using ML models that match customers to sales associates who best understand customer goals, preferences, and personalities. Implement advanced marketing mix models and strategies that enhance your CLV-to-CAC (customer acquisition cost) ratio. Deploy discount elasticity models that maximize margins, providing discounts only when and where most effective. The potential benefits are boundless.
The use cases for GenAI and ML are strong on their own, but together they are even stronger. Combining these tools effectively is a complex process—you may need to strengthen operational capabilities and bolster your AI and ML talent. But the payoff is worth it.
Integrating these two technologies into existing processes multiplies the result—often sooner and to a greater extent than expected. Start now and watch the effect accumulate.
