Bazooka Gum dumped ChatGPT for Claude, and it’s a perfect example of how old-school CPG brands are using AI to fight industry headwinds

(SeaPRwire) –   I caught up with Mark Henderson, a senior FMCG tech strategist with 18 years advising consumer goods brands on digital transformation, to get his take on Bazooka’s recent AI push. “I’ve watched CIOs get shuffled into back-office finance silos for decades, so seeing Bazooka elevate its tech lead to report directly to the CEO isn’t just an internal org chart tweak. It’s proof that for legacy consumer brands fighting for survival right now, AI isn’t a nice-to-have marketing gimmick, it’s the core of their cost-cutting and growth playbook. The switch from ChatGPT to Claude also tells me generic AI tools are already out of date for industry-specific use cases. Brands that don’t tailor their tech stacks to their actual operational pain points will get left behind fast.”

Sankar Karuppasamy knows those pain points firsthand. When he was CIO at trading card maker Topps, he reported to the CFO. After Bazooka split from Topps via a $700 million sale to Apax Partners in 2023, he stepped into the CIO role at the gum maker, and now reports directly to CEO Tony Jacobs. The shift lines up with a wider trend across US corporations, per Deloitte data, 65% of CIOs now report to the CEO, up from 41% a decade ago.

Bazooka has been around for 77 years, launching its signature gum in 1947, Bazooka Joe comics in 1953, and Ring Pops in 1977, but it’s facing unprecedented pressure right now. 2025 data from the National Confectioners Association shows US confectionary retail sales grew just 1.5% last year, with unit sales down across every category from chocolate to gum. Inflation is pushing consumers to cheaper private label options, and rising GLP-1 usage is cutting overall candy consumption across the board. Add to that the extreme seasonality of candy sales, which spike heavily around Halloween, Valentine’s Day and Christmas, and you get a perfect storm of operational challenges.

Karuppasamy’s first big win with AI was rolling out Anaplan’s ML-powered forecasting and scenario planning tools, which pushed demand forecasting accuracy up from 60% to 90% overnight, letting the team map manufacturing plans far more precisely to retailer demand. His wider AI playbook covers three core areas, general productivity gains, function-specific workflow overhauls, and early testing of agentic AI. The team initially used ChatGPT for general productivity use cases, but as they moved to build more custom, job-specific AI tools, they decided to shift the whole company to Anthropic’s Claude. Karuppasamy is open to switching again down the line, the AI space moves so fast that a better option could easily pop up next year.

The company is rolling out foundational LLM training for all employees, plus department-specific training focused on high-priority use cases, with AI adoption goals in place for both frontline staff and C-suite leaders. Use cases already live across the business include faster recipe development, pulling consumer insights from social media and retailer feedback, speeding up marketing material production while making sure packaging copy meets local regulatory requirements, and software upgrades for their Pennsylvania manufacturing facility. Karuppasamy has passed on adding robotics to the factory for now, given the relatively small size of the operation.

He’s also leading an agentic AI pilot focused on supply chain operations. The team is prioritizing building a clean, data-ready platform and strong governance guardrails to prevent data leaks before rolling out full agentic workflows, with initial testing focused on freight management and tariff analysis. Bazooka imports a large share of its products from across the globe, and manual tariff tracking has long been a costly, labor-intensive process for the team, the goal is to have AI handle that work autonomously, giving the finance team clear, real-time visibility into how tariff shifts will impact the company’s bottom line.

What’s playing out at Bazooka is far from an isolated case. Legacy CPG brands have lagged behind tech and e-commerce players on AI adoption for years, but the dual pressure of inflation and shifting consumer eating habits is forcing them to accelerate digital overhauls faster than anyone predicted. Over the next three years, we’ll see far more CIOs elevated to core leadership roles, as margin-strapped brands realize the cost savings from optimized forecasting and supply chain operations can easily add up to double-digit improvements to their bottom line. The shift from generic, off-the-shelf AI tools to model-specific, use case tailored setups will also pick up speed, as brands find generic tools can’t handle their unique operational needs, and fine-tuning options become more accessible and affordable. Agentic AI will likely become standard for FMCG supply chain teams within five years, cutting manual administrative work for logistics teams by 70% or more for brands that prioritize data quality and governance early on to avoid costly errors.