Can AI sketch a customer strategy before the data exists?

A team of second-year MBA students in the UB School of Management set out to test a different idea: what if AI could help sketch a strategy before all of that infrastructure is in place?

Adrija Aditya, Shirley Bassarath, Marissa Pecorella and Basil Zmiyiwsky explored that question through a hands-on marketing analytics project using Toasted, a Buffalo-based café brand, as their testing ground.

Instead of waiting for robust customer data, the student-led team, with oversight from their marketing professor, Wreetabrata Kar, built a synthetic dataset grounded in the business’s real-world context. That dataset became a kind of sandbox, providing a way to explore patterns, test ideas and pressure-check assumptions without the usual cost or delay.

From there, the students identified four distinct customer segments and evaluated three potential CRM approaches:

• Email campaigns to boost awareness and re-engage lapsed customers
• Loyalty enhancements to strengthen relationships with repeat visitors
• Personalized discounts to encourage short-term action

The results were telling. Loyalty-focused strategies showed the strongest potential, pointing to the value of deepening relationships with existing customers rather than leaning too heavily on promotions. In other words, consistent engagement may matter more than constant incentives.

There is a broader takeaway here. The project suggests that generative AI can act as an early testing layer, offering AI a way for businesses to explore strategy, simulate outcomes and make more informed decisions before investing in large-scale data collection or systems.

It is not a replacement for real customer insight. But it offers a starting point. And for smaller organizations trying to grow with limited resources, that starting point can make all the difference.

This story was written by AI and edited by a member of the UB School of Management Marketing and Communications Office.