Bazooka CIO deploys AI for demand planning and supply chain tasks

Sankar Karuppasamy has introduced machine learning tools at Bazooka to improve demand forecasting, speed up marketing material production, and support supply chain decisions. Karuppasamy previously served as chief information officer at trading card manufacturer Topps and reported to the chief financial officer.

After the 2023 separation of Bazooka from Topps through a $700 million sale to Apax Partners, he reports directly to CEO Tony Jacobs.

AI deployment Karuppasamy said the change reflects increased technology investment under the new owners. He has applied AI across three areas: general productivity, function-specific workflows, and an early pilot of agentic AI in supply chain operations.

Forecasting accuracy rose to 90 percent from 60 percent after Bazooka adopted Anaplan software that uses machine learning, Karuppasamy said. The company has also shifted internal use from OpenAI’s ChatGPT to Anthropic’s Claude for job-specific tasks.

Industry conditions U.S. retail sales of confectionery products grew 1.5 percent in 2025 while unit sales declined across chocolate, non-chocolate candy, gum, and mints, according to the National Confectioners Association. Bazooka faces additional pressure from seasonal demand spikes around Halloween, Valentine’s Day, and Christmas.

Karuppasamy has resisted adding robotics, citing the company’s smaller manufacturing footprint in Pennsylvania. A pilot program for agentic AI focuses first on data quality and governance before expanding to freight management and tariff analysis. Employees receive foundational training on large language models, and performance goals now include AI usage metrics for both staff and C-suite leaders.